Merge pull request #1038 from damemi/0.26.0-helm-chart

Update helm chart to v0.26.0

.github/ci/ct.yaml

@@ -0,0 +1,5 @@
+chart-dirs:
+  - charts
+helm-extra-args: "--timeout=5m"
+check-version-increment: false
+target-branch: master

.github/workflows/helm.yaml

@@ -0,0 +1,69 @@
+name: Helm
+
+on:
+  push:
+    branches:
+      - master
+      - release-*
+    paths:
+      - 'charts/**'
+      - '.github/workflows/helm.yaml'
+      - '.github/ci/ct.yaml'
+  pull_request:
+    paths:
+      - 'charts/**'
+      - '.github/workflows/helm.yaml'
+      - '.github/ci/ct.yaml'
+
+jobs:
+  lint-and-test:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Set up Helm
+        uses: azure/setup-helm@v2.1
+        with:
+          version: v3.9.2
+
+      - uses: actions/setup-python@v3.1.2
+        with:
+          python-version: 3.7
+
+      - uses: actions/setup-go@v3
+        with:
+          go-version: '1.19.3'
+
+      - name: Set up chart-testing
+        uses: helm/chart-testing-action@v2.2.1
+        with:
+          version: v3.7.0
+
+      - name: Run chart-testing (list-changed)
+        id: list-changed
+        run: |
+          changed=$(ct list-changed --config=.github/ci/ct.yaml)
+          if [[ -n "$changed" ]]; then
+            echo "::set-output name=changed::true"
+          fi
+
+      - name: Run chart-testing (lint)
+        run: ct lint --config=.github/ci/ct.yaml --validate-maintainers=false
+
+      # Need a multi node cluster so descheduler runs until evictions
+      - name: Create multi node Kind cluster
+        run: make kind-multi-node
+
+      # helm-extra-set-args only available after ct 3.6.0
+      - name: Run chart-testing (install)
+        run: ct install --config=.github/ci/ct.yaml --helm-extra-set-args='--set=kind=Deployment'
+
+      - name: E2E after chart install
+        env:
+          KUBERNETES_VERSION: "v1.26.0"
+          KIND_E2E: true
+          SKIP_INSTALL: true
+        run: make test-e2e

.github/workflows/security.yaml

@@ -0,0 +1,47 @@
+name: "Security"
+
+on:
+  push:
+    branches:
+      - main
+      - master
+      - release-*
+  schedule:
+    - cron: '30 1 * * 0'
+
+jobs:
+  analyze:
+    name: Analyze
+    runs-on: ubuntu-latest
+    permissions:
+      actions: read
+      contents: read
+      security-events: write
+
+    strategy:
+      fail-fast: false
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Build image
+        run: |
+          IMAGE_REPO=${HELM_IMAGE_REPO:-descheduler}
+          IMAGE_TAG=${HELM_IMAGE_TAG:-security-test}
+          VERSION=security-test make image
+      - name: Run Trivy vulnerability scanner
+        uses: aquasecurity/trivy-action@master
+        with:
+          image-ref: 'descheduler:security-test'
+          format: 'sarif'
+          exit-code: '0'
+          severity: 'CRITICAL,HIGH'
+          output: 'trivy-results.sarif'
+
+      - name: Upload Trivy scan results to GitHub Security tab
+        uses: github/codeql-action/upload-sarif@v2
+        with:
+          sarif_file: 'trivy-results.sarif'
+          exit-code: '0'

.gitignore

@@ -4,4 +4,5 @@ vendordiff.patch
 .idea/
 *.code-workspace
 .vscode/
-kind
+kind
+bin/

.golangci.yml

@@ -5,11 +5,14 @@ linters:
   disable-all: true
   enable:
     - gofmt
+    - gofumpt
     - gosimple
     - gocyclo
     - misspell
     - govet
 
 linters-settings:
+  gofumpt:
+    extra-rules: true
   goimports:
 local-prefixes: sigs.k8s.io/descheduler

Dockerfile

@@ -11,7 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-FROM golang:1.16.7
+FROM golang:1.19.3
 
 WORKDIR /go/src/sigs.k8s.io/descheduler
 COPY . .

Makefile

@@ -14,8 +14,10 @@
 
 .PHONY: test
 
+export CONTAINER_ENGINE ?= docker
+
 # VERSION is based on a date stamp plus the last commit
-VERSION?=v$(shell date +%Y%m%d)-$(shell git describe --tags --match "v*")
+VERSION?=v$(shell date +%Y%m%d)-$(shell git describe --tags)
 BRANCH?=$(shell git branch --show-current)
 SHA1?=$(shell git rev-parse HEAD)
 BUILD=$(shell date +%FT%T%z)
@@ -24,9 +26,12 @@ ARCHS = amd64 arm arm64
 
 LDFLAGS=-ldflags "-X ${LDFLAG_LOCATION}.version=${VERSION} -X ${LDFLAG_LOCATION}.buildDate=${BUILD} -X ${LDFLAG_LOCATION}.gitbranch=${BRANCH} -X ${LDFLAG_LOCATION}.gitsha1=${SHA1}"
 
-GOLANGCI_VERSION := v1.30.0
+GOLANGCI_VERSION := v1.49.0
 HAS_GOLANGCI := $(shell ls _output/bin/golangci-lint 2> /dev/null)
 
+GOFUMPT_VERSION := v0.4.0
+HAS_GOFUMPT := $(shell command -v gofumpt 2> /dev/null)
+
 # REGISTRY is the container registry to push
 # into. The default is to push to the staging
 # registry, not production.
@@ -60,36 +65,36 @@ build.arm64:
 	CGO_ENABLED=0 GOOS=linux GOARCH=arm64 go build ${LDFLAGS} -o _output/bin/descheduler sigs.k8s.io/descheduler/cmd/descheduler
 
 dev-image: build
-	docker build -f Dockerfile.dev -t $(IMAGE) .
+	$(CONTAINER_ENGINE) build -f Dockerfile.dev -t $(IMAGE) .
 
 image:
-	docker build --build-arg VERSION="$(VERSION)" --build-arg ARCH="amd64" -t $(IMAGE) .
+	$(CONTAINER_ENGINE) build --build-arg VERSION="$(VERSION)" --build-arg ARCH="amd64" -t $(IMAGE) .
 
 image.amd64:
-	docker build --build-arg VERSION="$(VERSION)" --build-arg ARCH="amd64" -t $(IMAGE)-amd64 .
+	$(CONTAINER_ENGINE) build --build-arg VERSION="$(VERSION)" --build-arg ARCH="amd64" -t $(IMAGE)-amd64 .
 
 image.arm:
-	docker build --build-arg VERSION="$(VERSION)" --build-arg ARCH="arm" -t $(IMAGE)-arm .
+	$(CONTAINER_ENGINE) build --build-arg VERSION="$(VERSION)" --build-arg ARCH="arm" -t $(IMAGE)-arm .
 
 image.arm64:
-	docker build --build-arg VERSION="$(VERSION)" --build-arg ARCH="arm64" -t $(IMAGE)-arm64 .
+	$(CONTAINER_ENGINE) build --build-arg VERSION="$(VERSION)" --build-arg ARCH="arm64" -t $(IMAGE)-arm64 .
 
 push: image
 	gcloud auth configure-docker
-	docker tag $(IMAGE) $(IMAGE_GCLOUD)
-	docker push $(IMAGE_GCLOUD)
+	$(CONTAINER_ENGINE) tag $(IMAGE) $(IMAGE_GCLOUD)
+	$(CONTAINER_ENGINE) push $(IMAGE_GCLOUD)
 
 push-all: image.amd64 image.arm image.arm64
 	gcloud auth configure-docker
 	for arch in $(ARCHS); do \
-		docker tag $(IMAGE)-$${arch} $(IMAGE_GCLOUD)-$${arch} ;\
-		docker push $(IMAGE_GCLOUD)-$${arch} ;\
+		$(CONTAINER_ENGINE) tag $(IMAGE)-$${arch} $(IMAGE_GCLOUD)-$${arch} ;\
+		$(CONTAINER_ENGINE) push $(IMAGE_GCLOUD)-$${arch} ;\
 	done
-	DOCKER_CLI_EXPERIMENTAL=enabled docker manifest create $(IMAGE_GCLOUD) $(addprefix --amend $(IMAGE_GCLOUD)-, $(ARCHS))
+	DOCKER_CLI_EXPERIMENTAL=enabled $(CONTAINER_ENGINE) manifest create $(IMAGE_GCLOUD) $(addprefix --amend $(IMAGE_GCLOUD)-, $(ARCHS))
 	for arch in $(ARCHS); do \
-		DOCKER_CLI_EXPERIMENTAL=enabled docker manifest annotate --arch $${arch} $(IMAGE_GCLOUD) $(IMAGE_GCLOUD)-$${arch} ;\
+		DOCKER_CLI_EXPERIMENTAL=enabled $(CONTAINER_ENGINE) manifest annotate --arch $${arch} $(IMAGE_GCLOUD) $(IMAGE_GCLOUD)-$${arch} ;\
 	done
-	DOCKER_CLI_EXPERIMENTAL=enabled docker manifest push $(IMAGE_GCLOUD) ;\
+	DOCKER_CLI_EXPERIMENTAL=enabled $(CONTAINER_ENGINE) manifest push $(IMAGE_GCLOUD) ;\
 
 clean:
 	rm -rf _output
@@ -131,17 +136,34 @@ verify-gen:
 
 lint:
 ifndef HAS_GOLANGCI
-	curl -sfL https://install.goreleaser.com/github.com/golangci/golangci-lint.sh | sh -s -- -b ./_output/bin ${GOLANGCI_VERSION}
+	curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b ./_output/bin ${GOLANGCI_VERSION}
 endif
 	./_output/bin/golangci-lint run
 
-lint-chart: ensure-helm-install
-	helm lint ./charts/descheduler
+fmt:
+ifndef HAS_GOFUMPT
+	go install mvdan.cc/gofumpt@${GOFUMPT_VERSION}
+endif
+	gofumpt -w -extra .
 
-test-helm: ensure-helm-install
-	./test/run-helm-tests.sh
+# helm
 
 ensure-helm-install:
 ifndef HAS_HELM
 	curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3 && chmod 700 ./get_helm.sh && ./get_helm.sh
-endif
+endif
+
+lint-chart: ensure-helm-install
+	helm lint ./charts/descheduler
+
+build-helm:
+	helm package ./charts/descheduler --dependency-update --destination ./bin/chart
+
+test-helm: ensure-helm-install
+	./test/run-helm-tests.sh
+
+kind-multi-node:
+	kind create cluster --name kind --config ./hack/kind_config.yaml --wait 2m
+
+ct-helm:
+	./hack/verify-chart.sh

OWNERS

@@ -2,14 +2,15 @@ approvers:
 - damemi
 - ingvagabund
 - seanmalloy
+- a7i
 reviewers:
-- aveshagarwal
-- k82cn
-- ravisantoshgudimetla
 - damemi
 - seanmalloy
 - ingvagabund
 - lixiang233
+- a7i
+- janeliul
+- knelasevero
 emeritus_approvers:
 - aveshagarwal
 - k82cn

README.md

@@ -1,6 +1,10 @@
 [![Go Report Card](https://goreportcard.com/badge/sigs.k8s.io/descheduler)](https://goreportcard.com/report/sigs.k8s.io/descheduler)
 ![Release Charts](https://github.com/kubernetes-sigs/descheduler/workflows/Release%20Charts/badge.svg)
 
+<p align="center">
+    <img src="assets/logo/descheduler-stacked-color.png" width="40%" align="center" alt="descheduler">
+</p>
+
 # Descheduler for Kubernetes
 
 Scheduling in Kubernetes is the process of binding pending pods to nodes, and is performed by
@@ -50,6 +54,8 @@ Table of Contents
   - [Node Fit filtering](#node-fit-filtering)
 - [Pod Evictions](#pod-evictions)
   - [Pod Disruption Budget (PDB)](#pod-disruption-budget-pdb)
+- [High Availability](#high-availability)
+  - [Configure HA Mode](#configure-ha-mode)
 - [Metrics](#metrics)
 - [Compatibility Matrix](#compatibility-matrix)
 - [Getting Involved and Contributing](#getting-involved-and-contributing)
@@ -99,21 +105,21 @@ The descheduler helm chart is also listed on the [artifact hub](https://artifact
 ### Install Using Kustomize
 
 You can use kustomize to install descheduler.
-See the [resources | Kustomize](https://kubectl.docs.kubernetes.io/references/kustomize/resource/) for detailed instructions.
+See the [resources | Kustomize](https://kubectl.docs.kubernetes.io/references/kustomize/cmd/build/) for detailed instructions.
 
 Run As A Job
 ```
-kustomize build 'github.com/kubernetes-sigs/descheduler/kubernetes/job?ref=v0.22.0' | kubectl apply -f -
+kustomize build 'github.com/kubernetes-sigs/descheduler/kubernetes/job?ref=v0.26.0' | kubectl apply -f -
 ```
 
 Run As A CronJob
 ```
-kustomize build 'github.com/kubernetes-sigs/descheduler/kubernetes/cronjob?ref=v0.22.0' | kubectl apply -f -
+kustomize build 'github.com/kubernetes-sigs/descheduler/kubernetes/cronjob?ref=v0.26.0' | kubectl apply -f -
 ```
 
 Run As A Deployment
 ```
-kustomize build 'github.com/kubernetes-sigs/descheduler/kubernetes/deployment?ref=v0.22.0' | kubectl apply -f -
+kustomize build 'github.com/kubernetes-sigs/descheduler/kubernetes/deployment?ref=v0.26.0' | kubectl apply -f -
 ```
 
 ## User Guide
@@ -132,6 +138,8 @@ The policy includes a common configuration that applies to all the strategies:
 | `evictSystemCriticalPods` | `false` | [Warning: Will evict Kubernetes system pods] allows eviction of pods with any priority, including system pods like kube-dns |
 | `ignorePvcPods` | `false` | set whether PVC pods should be evicted or ignored |
 | `maxNoOfPodsToEvictPerNode` | `nil` | maximum number of pods evicted from each node (summed through all strategies) |
+| `maxNoOfPodsToEvictPerNamespace` | `nil` | maximum number of pods evicted from each namespace (summed through all strategies) |
+| `evictFailedBarePods` | `false` | allow eviction of pods without owner references and in failed phase |
 
 As part of the policy, the parameters associated with each strategy can be configured.
 See each strategy for details on available parameters.
@@ -142,6 +150,7 @@ See each strategy for details on available parameters.
 apiVersion: "descheduler/v1alpha1"
 kind: "DeschedulerPolicy"
 nodeSelector: prod=dev
+evictFailedBarePods: false
 evictLocalStoragePods: true
 evictSystemCriticalPods: true
 maxNoOfPodsToEvictPerNode: 40
@@ -216,6 +225,17 @@ These thresholds, `thresholds` and `targetThresholds`, could be tuned as per you
 strategy evicts pods from `overutilized nodes` (those with usage above `targetThresholds`) to `underutilized nodes`
 (those with usage below `thresholds`), it will abort if any number of `underutilized nodes` or `overutilized nodes` is zero.
 
+Additionally, the strategy accepts a `useDeviationThresholds` parameter.
+If that parameter is set to `true`, the thresholds are considered as percentage deviations from mean resource usage.
+`thresholds` will be deducted from the mean among all nodes and `targetThresholds` will be added to the mean.
+A resource consumption above (resp. below) this window is considered as overutilization (resp. underutilization).
+
+**NOTE:** Node resource consumption is determined by the requests and limits of pods, not actual usage.
+This approach is chosen in order to maintain consistency with the kube-scheduler, which follows the same
+design for scheduling pods onto nodes. This means that resource usage as reported by Kubelet (or commands
+like `kubectl top`) may differ from the calculated consumption, due to these components reporting
+actual usage metrics. Implementing metrics-based descheduling is currently TODO for the project.
+
 **Parameters:**
 
 |Name|Type|
@@ -223,9 +243,11 @@ strategy evicts pods from `overutilized nodes` (those with usage above `targetTh
 |`thresholds`|map(string:int)|
 |`targetThresholds`|map(string:int)|
 |`numberOfNodes`|int|
+|`useDeviationThresholds`|bool|
 |`thresholdPriority`|int (see [priority filtering](#priority-filtering))|
 |`thresholdPriorityClassName`|string (see [priority filtering](#priority-filtering))|
 |`nodeFit`|bool (see [node fit filtering](#node-fit-filtering))|
+|`Namespaces`|(see [namespace filtering](#namespace-filtering))|
 
 **Example:**
 
@@ -263,10 +285,10 @@ under utilized frequently or for a short period of time. By default, `numberOfNo
 
 ### HighNodeUtilization
 
-This strategy finds nodes that are under utilized and evicts pods from the nodes in the hope that these pods will be 
-scheduled compactly into fewer nodes.  Used in conjunction with node auto-scaling, this strategy is intended to help 
+This strategy finds nodes that are under utilized and evicts pods from the nodes in the hope that these pods will be
+scheduled compactly into fewer nodes.  Used in conjunction with node auto-scaling, this strategy is intended to help
 trigger down scaling of under utilized nodes.
-This strategy **must** be used with the scheduler strategy `MostRequestedPriority`. The parameters of this strategy are 
+This strategy **must** be used with the scheduler scoring strategy `MostAllocated`. The parameters of this strategy are
 configured under `nodeResourceUtilizationThresholds`.
 
 The under utilization of nodes is determined by a configurable threshold `thresholds`. The threshold
@@ -283,6 +305,12 @@ strategy evicts pods from `underutilized nodes` (those with usage below `thresho
 so that they can be recreated in appropriately utilized nodes.
 The strategy will abort if any number of `underutilized nodes` or `appropriately utilized nodes` is zero.
 
+**NOTE:** Node resource consumption is determined by the requests and limits of pods, not actual usage.
+This approach is chosen in order to maintain consistency with the kube-scheduler, which follows the same
+design for scheduling pods onto nodes. This means that resource usage as reported by Kubelet (or commands
+like `kubectl top`) may differ from the calculated consumption, due to these components reporting
+actual usage metrics. Implementing metrics-based descheduling is currently TODO for the project.
+
 **Parameters:**
 
 |Name|Type|
@@ -292,6 +320,7 @@ The strategy will abort if any number of `underutilized nodes` or `appropriately
 |`thresholdPriority`|int (see [priority filtering](#priority-filtering))|
 |`thresholdPriorityClassName`|string (see [priority filtering](#priority-filtering))|
 |`nodeFit`|bool (see [node fit filtering](#node-fit-filtering))|
+|`Namespaces`|(see [namespace filtering](#namespace-filtering))|
 
 **Example:**
 
@@ -397,10 +426,17 @@ pod "podA" with a toleration to tolerate a taint ``key=value:NoSchedule`` schedu
 node. If the node's taint is subsequently updated/removed, taint is no longer satisfied by its pods' tolerations
 and will be evicted.
 
+Node taints can be excluded from consideration by specifying a list of excludedTaints. If a node taint key **or**
+key=value matches an excludedTaints entry, the taint will be ignored.
+
+For example, excludedTaints entry "dedicated" would match all taints with key "dedicated", regardless of value.
+excludedTaints entry "dedicated=special-user" would match taints with key "dedicated" and value "special-user".
+
 **Parameters:**
 
 |Name|Type|
 |---|---|
+|`excludedTaints`|list(string)|
 |`thresholdPriority`|int (see [priority filtering](#priority-filtering))|
 |`thresholdPriorityClassName`|string (see [priority filtering](#priority-filtering))|
 |`namespaces`|(see [namespace filtering](#namespace-filtering))|
@@ -415,6 +451,10 @@ kind: "DeschedulerPolicy"
 strategies:
   "RemovePodsViolatingNodeTaints":
     enabled: true
+    params:
+      excludedTaints:
+      - dedicated=special-user # exclude taints with key "dedicated" and value "special-user"
+      - reserved # exclude all taints with key "reserved"
 ````
 
 ### RemovePodsViolatingTopologySpreadConstraint
@@ -456,9 +496,9 @@ strategies:
 
 This strategy makes sure that pods having too many restarts are removed from nodes. For example a pod with EBS/PD that
 can't get the volume/disk attached to the instance, then the pod should be re-scheduled to other nodes. Its parameters
-include `podRestartThreshold`, which is the number of restarts at which a pod should be evicted, and `includingInitContainers`,
-which determines whether init container restarts should be factored into that calculation.
-|`labelSelector`|(see [label filtering](#label-filtering))|
+include `podRestartThreshold`, which is the number of restarts (summed over all eligible containers) at which a pod
+should be evicted, and `includingInitContainers`, which determines whether init container restarts should be factored
+into that calculation.
 
 **Parameters:**
 
@@ -469,6 +509,7 @@ which determines whether init container restarts should be factored into that ca
 |`thresholdPriority`|int (see [priority filtering](#priority-filtering))|
 |`thresholdPriorityClassName`|string (see [priority filtering](#priority-filtering))|
 |`namespaces`|(see [namespace filtering](#namespace-filtering))|
+|`labelSelector`|(see [label filtering](#label-filtering))|
 |`nodeFit`|bool (see [node fit filtering](#node-fit-filtering))|
 
 **Example:**
@@ -489,19 +530,24 @@ strategies:
 
 This strategy evicts pods that are older than `maxPodLifeTimeSeconds`.
 
-You can also specify `podStatusPhases` to `only` evict pods with specific `StatusPhases`, currently this parameter is limited
-to `Running` and `Pending`.
+You can also specify `states` parameter to **only** evict pods matching the following conditions:
+  - [Pod Phase](https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/#pod-phase) status of: `Running`, `Pending`
+  - [Container State Waiting](https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/#container-state-waiting) condition of: `PodInitializing`, `ContainerCreating`
+
+If a value for `states` or `podStatusPhases` is not specified,
+Pods in any state (even `Running`) are considered for eviction.
 
 **Parameters:**
 
-|Name|Type|
-|---|---|
-|`maxPodLifeTimeSeconds`|int|
-|`podStatusPhases`|list(string)|
-|`thresholdPriority`|int (see [priority filtering](#priority-filtering))|
-|`thresholdPriorityClassName`|string (see [priority filtering](#priority-filtering))|
-|`namespaces`|(see [namespace filtering](#namespace-filtering))|
-|`labelSelector`|(see [label filtering](#label-filtering))|
+|Name|Type|Notes|
+|---|---|---|
+|`maxPodLifeTimeSeconds`|int||
+|`podStatusPhases`|list(string)|Deprecated in v0.25+ Use `states` instead|
+|`states`|list(string)|Only supported in v0.25+|
+|`thresholdPriority`|int (see [priority filtering](#priority-filtering))||
+|`thresholdPriorityClassName`|string (see [priority filtering](#priority-filtering))||
+|`namespaces`|(see [namespace filtering](#namespace-filtering))||
+|`labelSelector`|(see [label filtering](#label-filtering))||
 
 **Example:**
 
@@ -514,8 +560,9 @@ strategies:
      params:
        podLifeTime:
          maxPodLifeTimeSeconds: 86400
-         podStatusPhases:
+         states:
          - "Pending"
+         - "PodInitializing"
 ```
 
 ### RemoveFailedPods
@@ -523,7 +570,7 @@ strategies:
 This strategy evicts pods that are in failed status phase.
 You can provide an optional parameter to filter by failed `reasons`.
 `reasons` can be expanded to include reasons of InitContainers as well by setting the optional parameter `includingInitContainers` to `true`.
-You can specify an optional parameter `minPodLifeTimeSeconds` to evict pods that are older than specified seconds.
+You can specify an optional parameter `minPodLifetimeSeconds` to evict pods that are older than specified seconds.
 Lastly, you can specify the optional parameter `excludeOwnerKinds` and if a pod
 has any of these `Kind`s listed as an `OwnerRef`, that pod will not be considered for eviction.
 
@@ -531,7 +578,7 @@ has any of these `Kind`s listed as an `OwnerRef`, that pod will not be considere
 
 |Name|Type|
 |---|---|
-|`minPodLifeTimeSeconds`|uint|
+|`minPodLifetimeSeconds`|uint|
 |`excludeOwnerKinds`|list(string)|
 |`reasons`|list(string)|
 |`includingInitContainers`|bool|
@@ -556,7 +603,7 @@ strategies:
          includingInitContainers: true
          excludeOwnerKinds:
          - "Job"
-         minPodLifeTimeSeconds: 3600
+         minPodLifetimeSeconds: 3600
 ```
 
 ## Filter Pods
@@ -572,6 +619,7 @@ The following strategies accept a `namespaces` parameter which allows to specify
 * `RemoveDuplicates`
 * `RemovePodsViolatingTopologySpreadConstraint`
 * `RemoveFailedPods`
+* `LowNodeUtilization` and `HighNodeUtilization` (Only filtered right before eviction)
 
 For example:
 
@@ -654,7 +702,7 @@ does not exist, descheduler won't create it and will throw an error.
 
 ### Label filtering
 
-The following strategies can configure a [standard kubernetes labelSelector](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.22/#labelselector-v1-meta)
+The following strategies can configure a [standard kubernetes labelSelector](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#labelselector-v1-meta)
 to filter pods by their labels:
 
 * `PodLifeTime`
@@ -702,8 +750,9 @@ The following strategies accept a `nodeFit` boolean parameter which can optimize
 
  If set to `true` the descheduler will consider whether or not the pods that meet eviction criteria will fit on other nodes before evicting them. If a pod cannot be rescheduled to another node, it will not be evicted. Currently the following criteria are considered when setting `nodeFit` to `true`:
 - A `nodeSelector` on the pod
-- Any `Tolerations` on the pod and any `Taints` on the other nodes
+- Any `tolerations` on the pod and any `taints` on the other nodes
 - `nodeAffinity` on the pod
+- Resource `requests` made by the pod and the resources available on other nodes
 - Whether any of the other nodes are marked as `unschedulable`
 
 E.g.
@@ -713,18 +762,18 @@ apiVersion: "descheduler/v1alpha1"
 kind: "DeschedulerPolicy"
 strategies:
   "LowNodeUtilization":
-     enabled: true
-     params:
-       nodeResourceUtilizationThresholds:
-         thresholds:
-           "cpu" : 20
-           "memory": 20
-           "pods": 20
-         targetThresholds:
-           "cpu" : 50
-           "memory": 50
-           "pods": 50
-        nodeFit: true
+    enabled: true
+    params:
+      nodeFit: true
+      nodeResourceUtilizationThresholds:
+        thresholds:
+          "cpu": 20
+          "memory": 20
+          "pods": 20
+        targetThresholds:
+          "cpu": 50
+          "memory": 50
+          "pods": 50
 ```
 
 Note that node fit filtering references the current pod spec, and not that of it's owner.
@@ -739,8 +788,8 @@ Using Deployments instead of ReplicationControllers provides an automated rollou
 When the descheduler decides to evict pods from a node, it employs the following general mechanism:
 
 * [Critical pods](https://kubernetes.io/docs/tasks/administer-cluster/guaranteed-scheduling-critical-addon-pods/) (with priorityClassName set to system-cluster-critical or system-node-critical) are never evicted (unless `evictSystemCriticalPods: true` is set).
-* Pods (static or mirrored pods or stand alone pods) not part of an ReplicationController, ReplicaSet(Deployment), StatefulSet, or Job are
-never evicted because these pods won't be recreated.
+* Pods (static or mirrored pods or standalone pods) not part of an ReplicationController, ReplicaSet(Deployment), StatefulSet, or Job are
+never evicted because these pods won't be recreated. (Standalone pods in failed status phase can be evicted by setting `evictFailedBarePods: true`)
 * Pods associated with DaemonSets are never evicted.
 * Pods with local storage are never evicted (unless `evictLocalStoragePods: true` is set).
 * Pods with PVCs are evicted (unless `ignorePvcPods: true` is set).
@@ -749,6 +798,10 @@ best effort pods are evicted before burstable and guaranteed pods.
 * All types of pods with the annotation `descheduler.alpha.kubernetes.io/evict` are eligible for eviction. This
   annotation is used to override checks which prevent eviction and users can select which pod is evicted.
   Users should know how and if the pod will be recreated.
+  The annotation only affects internal descheduler checks.
+  The anti-disruption protection provided by the [/eviction](https://kubernetes.io/docs/concepts/scheduling-eviction/api-eviction/)
+  subresource is still respected.
+* Pods with a non-nil DeletionTimestamp are not evicted by default.
 
 Setting `--v=4` or greater on the Descheduler will log all reasons why any pod is not evictable.
 
@@ -757,6 +810,23 @@ Setting `--v=4` or greater on the Descheduler will log all reasons why any pod i
 Pods subject to a Pod Disruption Budget(PDB) are not evicted if descheduling violates its PDB. The pods
 are evicted by using the eviction subresource to handle PDB.
 
+## High Availability
+
+In High Availability mode, Descheduler starts [leader election](https://github.com/kubernetes/client-go/tree/master/tools/leaderelection) process in Kubernetes. You can activate HA mode
+if you choose to deploy your application as Deployment.
+
+Deployment starts with 1 replica by default. If you want to use more than 1 replica, you must consider
+enable High Availability mode since we don't want to run descheduler pods simultaneously.
+
+### Configure HA Mode
+
+The leader election process can be enabled by setting `--leader-elect` in the CLI. You can also set
+`--set=leaderElection.enabled=true` flag if you are using Helm.
+
+To get best results from HA mode some additional configurations might require:
+* Configure a [podAntiAffinity](https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node) rule if you want to schedule onto a node only if that node is in the same zone as at least one already-running descheduler
+* Set the replica count greater than 1
+
 ## Metrics
 
 | name	| type	| description |
@@ -776,17 +846,20 @@ v0.18 should work with k8s v1.18, v1.17, and v1.16.
 Starting with descheduler release v0.18 the minor version of descheduler matches the minor version of the k8s client
 packages that it is compiled with.
 
-Descheduler  | Supported Kubernetes Version
--------------|-----------------------------
-v0.22        | v1.22
-v0.21        | v1.21
-v0.20        | v1.20
-v0.19        | v1.19
-v0.18        | v1.18
-v0.10        | v1.17
-v0.4-v0.9    | v1.9+
-v0.1-v0.3    | v1.7-v1.8
-
+| Descheduler | Supported Kubernetes Version |
+|-------------|------------------------------|
+| v0.26       | v1.26                        |
+| v0.25       | v1.25                        |
+| v0.24       | v1.24                        |
+| v0.23       | v1.23                        |
+| v0.22       | v1.22                        |
+| v0.21       | v1.21                        |
+| v0.20       | v1.20                        |
+| v0.19       | v1.19                        |
+| v0.18       | v1.18                        |
+| v0.10       | v1.17                        |
+| v0.4-v0.9   | v1.9+                        |
+| v0.1-v0.3   | v1.7-v1.8                    |
 
 ## Getting Involved and Contributing
 

charts/descheduler/Chart.yaml

@@ -1,7 +1,7 @@
 apiVersion: v1
 name: descheduler
-version: 0.22.0
-appVersion: 0.22.0
+version: 0.26.0
+appVersion: 0.26.0
 description: Descheduler for Kubernetes is used to rebalance clusters by evicting pods that can potentially be scheduled on better nodes. In the current implementation, descheduler does not schedule replacement of evicted pods but relies on the default scheduler for that.
 keywords:
 - kubernetes

charts/descheduler/README.md

@@ -43,28 +43,46 @@ The command removes all the Kubernetes components associated with the chart and
 
 The following table lists the configurable parameters of the _descheduler_ chart and their default values.
 
-| Parameter                      | Description                                                                                                           | Default                              |
-| ------------------------------ | --------------------------------------------------------------------------------------------------------------------- | ------------------------------------ |
-| `kind`                         | Use as CronJob or Deployment                                                                                          | `CronJob`                            |
-| `image.repository`             | Docker repository to use                                                                                              | `k8s.gcr.io/descheduler/descheduler` |
-| `image.tag`                    | Docker tag to use                                                                                                     | `v[chart appVersion]`                |
-| `image.pullPolicy`             | Docker image pull policy                                                                                              | `IfNotPresent`                       |
-| `imagePullSecrets`             | Docker repository secrets                                                                                              | `[]`                       |
-| `nameOverride`                 | String to partially override `descheduler.fullname` template (will prepend the release name)                          | `""`                                 |
-| `fullnameOverride`             | String to fully override `descheduler.fullname` template                                                              | `""`                                 |
-| `cronJobApiVersion`            | CronJob API Group Version                                                                                             | `"batch/v1"`                         |
-| `schedule`                     | The cron schedule to run the _descheduler_ job on                                                                     | `"*/2 * * * *"`                      |
-| `startingDeadlineSeconds`      | If set, configure `startingDeadlineSeconds` for the _descheduler_ job                                                 | `nil`                                |
-| `successfulJobsHistoryLimit`   | If set, configure `successfulJobsHistoryLimit` for the _descheduler_ job                                              | `nil`                                |
-| `failedJobsHistoryLimit`       | If set, configure `failedJobsHistoryLimit` for the _descheduler_ job                                                  | `nil`                                |
-| `deschedulingInterval`         | If using kind:Deployment, sets time between consecutive descheduler executions.                                       | `5m`                                 |
-| `cmdOptions`                   | The options to pass to the _descheduler_ command                                                                      | _see values.yaml_                    |
-| `deschedulerPolicy.strategies` | The _descheduler_ strategies to apply                                                                                 | _see values.yaml_                    |
-| `priorityClassName`            | The name of the priority class to add to pods                                                                         | `system-cluster-critical`            |
-| `rbac.create`                  | If `true`, create & use RBAC resources                                                                                | `true`                               |
-| `podSecurityPolicy.create`     | If `true`, create PodSecurityPolicy                                                                                   | `true`                               |
-| `resources`                    | Descheduler container CPU and memory requests/limits                                                                  | _see values.yaml_                    |
-| `serviceAccount.create`        | If `true`, create a service account for the cron job                                                                  | `true`                               |
-| `serviceAccount.name`          | The name of the service account to use, if not set and create is true a name is generated using the fullname template | `nil`                                |
-| `nodeSelector`                 | Node selectors to run the descheduler cronjob on specific nodes                                                       | `nil`                                |
-| `tolerations`                  | tolerations to run the descheduler cronjob on specific nodes                                                          | `nil`                                |
+| Parameter                           | Description                                                                                                           | Default                                   |
+| ----------------------------------- | --------------------------------------------------------------------------------------------------------------------- | ----------------------------------------- |
+| `kind`                              | Use as CronJob or Deployment                                                                                          | `CronJob`                                 |
+| `image.repository`                  | Docker repository to use                                                                                              | `registry.k8s.io/descheduler/descheduler` |
+| `image.tag`                         | Docker tag to use                                                                                                     | `v[chart appVersion]`                     |
+| `image.pullPolicy`                  | Docker image pull policy                                                                                              | `IfNotPresent`                            |
+| `imagePullSecrets`                  | Docker repository secrets                                                                                             | `[]`                                      |
+| `nameOverride`                      | String to partially override `descheduler.fullname` template (will prepend the release name)                          | `""`                                      |
+| `fullnameOverride`                  | String to fully override `descheduler.fullname` template                                                              | `""`                                      |
+| `cronJobApiVersion`                 | CronJob API Group Version                                                                                             | `"batch/v1"`                              |
+| `schedule`                          | The cron schedule to run the _descheduler_ job on                                                                     | `"*/2 * * * *"`                           |
+| `startingDeadlineSeconds`           | If set, configure `startingDeadlineSeconds` for the _descheduler_ job                                                 | `nil`                                     |
+| `successfulJobsHistoryLimit`        | If set, configure `successfulJobsHistoryLimit` for the _descheduler_ job                                              | `3`                                       |
+| `failedJobsHistoryLimit`            | If set, configure `failedJobsHistoryLimit` for the _descheduler_ job                                                  | `1`                                       |
+| `ttlSecondsAfterFinished`           | If set, configure `ttlSecondsAfterFinished` for the _descheduler_ job                                                 | `nil`                                     |
+| `deschedulingInterval`              | If using kind:Deployment, sets time between consecutive descheduler executions.                                       | `5m`                                      |
+| `replicas`                          | The replica count for Deployment                                                                                      | `1`                                       |
+| `leaderElection`                    | The options for high availability when running replicated components                                                  | _see values.yaml_                         |
+| `cmdOptions`                        | The options to pass to the _descheduler_ command                                                                      | _see values.yaml_                         |
+| `deschedulerPolicy.strategies`      | The _descheduler_ strategies to apply                                                                                 | _see values.yaml_                         |
+| `priorityClassName`                 | The name of the priority class to add to pods                                                                         | `system-cluster-critical`                 |
+| `rbac.create`                       | If `true`, create & use RBAC resources                                                                                | `true`                                    |
+| `resources`                         | Descheduler container CPU and memory requests/limits                                                                  | _see values.yaml_                         |
+| `serviceAccount.create`             | If `true`, create a service account for the cron job                                                                  | `true`                                    |
+| `serviceAccount.name`               | The name of the service account to use, if not set and create is true a name is generated using the fullname template | `nil`                                     |
+| `serviceAccount.annotations`        | Specifies custom annotations for the serviceAccount                                                                   | `{}`                                      |
+| `podAnnotations`                    | Annotations to add to the descheduler Pods                                                                            | `{}`                                      |
+| `podLabels`                         | Labels to add to the descheduler Pods                                                                                 | `{}`                                      |
+| `nodeSelector`                      | Node selectors to run the descheduler cronjob/deployment on specific nodes                                            | `nil`                                     |
+| `service.enabled`                   | If `true`, create a service for deployment                                                                            | `false`                                   |
+| `serviceMonitor.enabled`            | If `true`, create a ServiceMonitor for deployment                                                                     | `false`                                   |
+| `serviceMonitor.namespace`          | The namespace where Prometheus expects to find service monitors                                                       | `nil`                                     |
+| `serviceMonitor.interval`           | The scrape interval. If not set, the Prometheus default scrape interval is used                                       | `nil`                                     |
+| `serviceMonitor.honorLabels`        | Keeps the scraped data's labels when labels are on collisions with target labels.                                     | `true`                                    |
+| `serviceMonitor.insecureSkipVerify` | Skip TLS certificate validation when scraping                                                                         | `true`                                    |
+| `serviceMonitor.serverName`         | Name of the server to use when validating TLS certificate                                                             | `nil`                                     |
+| `serviceMonitor.metricRelabelings`  | MetricRelabelConfigs to apply to samples after scraping, but before ingestion                                         | `[]`                                      |
+| `serviceMonitor.relabelings`        | RelabelConfigs to apply to samples before scraping                                                                    | `[]`                                      |
+| `affinity`                          | Node affinity to run the descheduler cronjob/deployment on specific nodes                                             | `nil`                                     |
+| `tolerations`                       | tolerations to run the descheduler cronjob/deployment on specific nodes                                               | `nil`                                     |
+| `suspend`                           | Set spec.suspend in descheduler cronjob                                                                               | `false`                                   |
+| `commonLabels`                      | Labels to apply to all resources                                                                                      | `{}`                                      |
+| `livenessProbe`                     | Liveness probe configuration for the descheduler container                                                            | _see values.yaml_                         |

charts/descheduler/templates/NOTES.txt

@@ -1 +1,12 @@
-Descheduler installed as a cron job.
+Descheduler installed as a {{ .Values.kind }}.
+
+{{- if eq .Values.kind "Deployment" }}
+{{- if eq .Values.replicas 1.0}}
+WARNING: You set replica count as 1 and workload kind as Deployment however leaderElection is not enabled. Consider enabling Leader Election for HA mode.
+{{- end}}
+{{- if .Values.leaderElection }}
+{{- if and (hasKey .Values.cmdOptions "dry-run") (eq (get .Values.cmdOptions "dry-run") true) }}
+WARNING: You enabled DryRun mode, you can't use Leader Election.
+{{- end}}
+{{- end}}
+{{- end}}

charts/descheduler/templates/_helpers.tpl

@@ -42,6 +42,9 @@ app.kubernetes.io/instance: {{ .Release.Name }}
 app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
 {{- end }}
 app.kubernetes.io/managed-by: {{ .Release.Service }}
+{{- if .Values.commonLabels}}
+{{ toYaml .Values.commonLabels }}
+{{- end }}
 {{- end -}}
 
 {{/*
@@ -62,3 +65,30 @@ Create the name of the service account to use
     {{ default "default" .Values.serviceAccount.name }}
 {{- end -}}
 {{- end -}}
+
+{{/*
+Leader Election
+*/}}
+{{- define "descheduler.leaderElection"}}
+{{- if .Values.leaderElection -}}
+- --leader-elect={{ .Values.leaderElection.enabled }}
+{{- if .Values.leaderElection.leaseDuration }}
+- --leader-elect-lease-duration={{ .Values.leaderElection.leaseDuration }}
+{{- end }}
+{{- if .Values.leaderElection.renewDeadline }}
+- --leader-elect-renew-deadline={{ .Values.leaderElection.renewDeadline }}
+{{- end }}
+{{- if .Values.leaderElection.retryPeriod }}
+- --leader-elect-retry-period={{ .Values.leaderElection.retryPeriod }}
+{{- end }}
+{{- if .Values.leaderElection.resourceLock }}
+- --leader-elect-resource-lock={{ .Values.leaderElection.resourceLock }}
+{{- end }}
+{{- if .Values.leaderElection.resourceName }}
+- --leader-elect-resource-name={{ .Values.leaderElection.resourceName }}
+{{- end }}
+{{- if .Values.leaderElection.resourceNamescape }}
+- --leader-elect-resource-namespace={{ .Values.leaderElection.resourceNamescape }}
+{{- end -}}
+{{- end }}
+{{- end }}

charts/descheduler/templates/clusterrole.yaml

@@ -6,7 +6,7 @@ metadata:
   labels:
     {{- include "descheduler.labels" . | nindent 4 }}
 rules:
-- apiGroups: [""]
+- apiGroups: ["events.k8s.io"]
   resources: ["events"]
   verbs: ["create", "update"]
 - apiGroups: [""]
@@ -14,7 +14,7 @@ rules:
   verbs: ["get", "watch", "list"]
 - apiGroups: [""]
   resources: ["namespaces"]
-  verbs: ["get", "list"]
+  verbs: ["get", "watch", "list"]
 - apiGroups: [""]
   resources: ["pods"]
   verbs: ["get", "watch", "list", "delete"]
@@ -24,11 +24,13 @@ rules:
 - apiGroups: ["scheduling.k8s.io"]
   resources: ["priorityclasses"]
   verbs: ["get", "watch", "list"]
-{{- if .Values.podSecurityPolicy.create }}
-- apiGroups: ['policy']
-  resources: ['podsecuritypolicies']
-  verbs:     ['use']
-  resourceNames:
-  - {{ template "descheduler.fullname" . }}
+{{- if .Values.leaderElection.enabled }}
+- apiGroups: ["coordination.k8s.io"]
+  resources: ["leases"]
+  verbs: ["create", "update"]
+- apiGroups: ["coordination.k8s.io"]
+  resources: ["leases"]
+  resourceNames: ["{{ .Values.leaderElection.resourceName | default "descheduler" }}"]
+  verbs: ["get", "patch", "delete"]
 {{- end }}
 {{- end -}}

charts/descheduler/templates/configmap.yaml

@@ -2,6 +2,7 @@ apiVersion: v1
 kind: ConfigMap
 metadata:
   name: {{ template "descheduler.fullname" . }}
+  namespace: {{ .Release.Namespace }}
   labels:
     {{- include "descheduler.labels" . | nindent 4 }}
 data:

charts/descheduler/templates/cronjob.yaml

@@ -3,10 +3,14 @@ apiVersion: {{ .Values.cronJobApiVersion | default "batch/v1" }}
 kind: CronJob
 metadata:
   name: {{ template "descheduler.fullname" . }}
+  namespace: {{ .Release.Namespace }}
   labels:
     {{- include "descheduler.labels" . | nindent 4 }}
 spec:
   schedule: {{ .Values.schedule | quote }}
+  {{- if .Values.suspend }}
+  suspend: {{ .Values.suspend }}
+  {{- end }}
   concurrencyPolicy: "Forbid"
   {{- if .Values.startingDeadlineSeconds }}
   startingDeadlineSeconds: {{ .Values.startingDeadlineSeconds }}
@@ -19,6 +23,9 @@ spec:
   {{- end }}
   jobTemplate:
     spec:
+      {{- if .Values.ttlSecondsAfterFinished }}
+      ttlSecondsAfterFinished: {{ .Values.ttlSecondsAfterFinished }}
+      {{- end }}
       template:
         metadata:
           name: {{ template "descheduler.fullname" . }}
@@ -37,6 +44,10 @@ spec:
           nodeSelector:
             {{- toYaml . | nindent 12 }}
           {{- end }}
+          {{- with .Values.affinity }}
+          affinity:
+            {{- toYaml . | nindent 12 }}
+          {{- end }}
           {{- with .Values.tolerations }}
           tolerations:
             {{- toYaml . | nindent 12 }}
@@ -65,6 +76,8 @@ spec:
                 - {{ $value | quote }}
                 {{- end }}
                 {{- end }}
+              livenessProbe:
+                {{- toYaml .Values.livenessProbe | nindent 16 }}
               resources:
                 {{- toYaml .Values.resources | nindent 16 }}
               securityContext:

charts/descheduler/templates/deployment.yaml

@@ -3,10 +3,18 @@ apiVersion: apps/v1
 kind: Deployment
 metadata:
   name: {{ template "descheduler.fullname" . }}
+  namespace: {{ .Release.Namespace }}
   labels:
     {{- include "descheduler.labels" . | nindent 4 }}
 spec:
+  {{- if gt .Values.replicas 1.0}}
+  {{- if not .Values.leaderElection.enabled }}
+  {{- fail "You must set leaderElection to use more than 1 replica"}}
+  {{- end}}
+  replicas: {{ required "leaderElection required for running more than one replica" .Values.replicas }}
+  {{- else }}
   replicas: 1
+  {{- end }}
   selector:
     matchLabels:
       {{- include "descheduler.selectorLabels" . | nindent 6 }}
@@ -27,6 +35,10 @@ spec:
       priorityClassName: {{ .Values.priorityClassName }}
       {{- end }}
       serviceAccountName: {{ template "descheduler.serviceAccountName" . }}
+      {{- with .Values.imagePullSecrets }}
+      imagePullSecrets:
+      {{- toYaml . | nindent 10 }}
+      {{- end }}
       containers:
         - name: {{ .Chart.Name }}
           image: "{{ .Values.image.repository }}:{{ .Values.image.tag | default (printf "v%s" .Chart.AppVersion) }}"
@@ -44,9 +56,12 @@ spec:
             - {{ $value | quote }}
             {{- end }}
             {{- end }}
+            {{- include "descheduler.leaderElection" . | nindent 12 }}
           ports:
             - containerPort: 10258
               protocol: TCP
+          livenessProbe:
+            {{- toYaml .Values.livenessProbe | nindent 12 }}
           resources:
             {{- toYaml .Values.resources | nindent 12 }}
           securityContext:
@@ -68,6 +83,10 @@ spec:
       nodeSelector:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.affinity }}
+      affinity:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
       {{- with .Values.tolerations }}
       tolerations:
         {{- toYaml . | nindent 8 }}

charts/descheduler/templates/podsecuritypolicy.yaml

@@ -1,38 +0,0 @@
-{{- if .Values.podSecurityPolicy.create -}}
-apiVersion: policy/v1beta1
-kind: PodSecurityPolicy
-metadata:
-  name: {{ template "descheduler.fullname" . }}
-  annotations:
-    seccomp.security.alpha.kubernetes.io/allowedProfileNames: 'docker/default,runtime/default'
-    seccomp.security.alpha.kubernetes.io/defaultProfileName:  'runtime/default'
-spec:
-  privileged: false
-  allowPrivilegeEscalation: false
-  requiredDropCapabilities:
-    - ALL
-  volumes:
-    - 'configMap'
-    - 'secret'
-  hostNetwork: false
-  hostIPC: false
-  hostPID: false
-  runAsUser:
-    rule: 'MustRunAs'
-    ranges:
-      - min: 1
-        max: 65535
-  seLinux:
-    rule: 'RunAsAny'
-  supplementalGroups:
-    rule: 'MustRunAs'
-    ranges:
-      - min: 1
-        max: 65535
-  fsGroup:
-    rule: 'MustRunAs'
-    ranges:
-      - min: 1
-        max: 65535
-  readOnlyRootFilesystem: true
-{{- end -}}

charts/descheduler/templates/service.yaml

@@ -0,0 +1,21 @@
+{{- if eq .Values.kind "Deployment" }}
+{{- if eq .Values.service.enabled true }}
+apiVersion: v1
+kind: Service
+metadata:
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+  name: {{ template "descheduler.fullname" . }}
+  namespace: {{ .Release.Namespace }}
+spec:
+  clusterIP: None
+  ports:
+  - name: http-metrics
+    port: 10258
+    protocol: TCP
+    targetPort: 10258
+  selector:
+    {{- include "descheduler.selectorLabels" . | nindent 4 }}
+  type: ClusterIP
+{{- end }}
+{{- end }}

charts/descheduler/templates/serviceaccount.yaml

@@ -3,6 +3,10 @@ apiVersion: v1
 kind: ServiceAccount
 metadata:
   name: {{ template "descheduler.serviceAccountName" . }}
+  namespace: {{ .Release.Namespace }}
   labels:
     {{- include "descheduler.labels" . | nindent 4 }}
+{{- if .Values.serviceAccount.annotations }}
+  annotations: {{ toYaml .Values.serviceAccount.annotations | nindent 4 }}
+{{- end }}
 {{- end -}}

charts/descheduler/templates/servicemonitor.yaml

@@ -0,0 +1,41 @@
+{{- if eq .Values.kind "Deployment" }}
+{{- if eq .Values.serviceMonitor.enabled true }}
+apiVersion: monitoring.coreos.com/v1
+kind: ServiceMonitor
+metadata:
+  name: {{ template "descheduler.fullname" . }}-servicemonitor
+  namespace: {{ .Values.serviceMonitor.namespace | default .Release.Namespace }}
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+spec:
+  jobLabel: jobLabel
+  namespaceSelector:
+    matchNames:
+    - {{ .Release.Namespace }}
+  selector:
+    matchLabels:
+      {{- include "descheduler.selectorLabels" . | nindent 6 }}
+  endpoints:
+  - honorLabels: {{ .Values.serviceMonitor.honorLabels | default true }}
+    port: http-metrics
+    {{- if .Values.serviceMonitor.interval }}
+    interval: {{ .Values.serviceMonitor.interval }}
+    {{- end }}
+    scheme: https
+    tlsConfig:
+      {{- if eq .Values.serviceMonitor.insecureSkipVerify true }}
+      insecureSkipVerify: true
+      {{- end }}
+      {{- if .Values.serviceMonitor.serverName }}
+      serverName: {{ .Values.serviceMonitor.serverName }}
+      {{- end}}
+{{- if .Values.serviceMonitor.metricRelabelings }}
+    metricRelabelings:
+{{ tpl (toYaml .Values.serviceMonitor.metricRelabelings | indent 4) . }}
+{{- end }}
+{{- if .Values.serviceMonitor.relabelings }}
+    relabelings:
+{{ tpl (toYaml .Values.serviceMonitor.relabelings | indent 4) . }}
+{{- end }}
+{{- end }}
+{{- end }}

charts/descheduler/templates/tests/test-descheduler-pod.yaml

@@ -1,29 +0,0 @@
-apiVersion: v1
-kind: Pod
-metadata:
-    name: descheduler-test-pod
-    annotations:
-      "helm.sh/hook": test
-spec:
-    restartPolicy: Never
-    serviceAccountName: descheduler-ci
-    containers:
-      - name: descheduler-test-container
-        image: alpine:latest
-        imagePullPolicy: IfNotPresent
-        securityContext:
-            allowPrivilegeEscalation: false
-            capabilities:
-              drop:
-                - All
-            privileged: false
-            runAsNonRoot: false 
-        command: ["/bin/ash"]
-        args:
-          - -c
-          - >-
-            apk --no-cache add curl &&
-            curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/linux/amd64/kubectl &&
-            chmod +x ./kubectl &&
-            mv ./kubectl /usr/local/bin/kubectl &&
-            /usr/local/bin/kubectl get pods --namespace kube-system --token "$(cat /var/run/secrets/kubernetes.io/serviceaccount/token)" | grep "descheduler" | grep "Completed"

charts/descheduler/values.yaml

@@ -6,12 +6,13 @@
 kind: CronJob
 
 image:
-  repository: k8s.gcr.io/descheduler/descheduler
+  repository: registry.k8s.io/descheduler/descheduler
   # Overrides the image tag whose default is the chart version
   tag: ""
   pullPolicy: IfNotPresent
 
-imagePullSecrets: []
+imagePullSecrets:
+#   - name: container-registry-secret
 
 resources:
   requests:
@@ -24,25 +25,56 @@ resources:
 nameOverride: ""
 fullnameOverride: ""
 
-cronJobApiVersion: "batch/v1"  # Use "batch/v1beta1" for k8s version < 1.21.0. TODO(@7i) remove with 1.23 release
+# labels that'll be applied to all resources
+commonLabels: {}
+
+cronJobApiVersion: "batch/v1"
 schedule: "*/2 * * * *"
-#startingDeadlineSeconds: 200
-#successfulJobsHistoryLimit: 1
-#failedJobsHistoryLimit: 1
+suspend: false
+# startingDeadlineSeconds: 200
+# successfulJobsHistoryLimit: 3
+# failedJobsHistoryLimit: 1
+# ttlSecondsAfterFinished 600
 
 # Required when running as a Deployment
 deschedulingInterval: 5m
 
+# Specifies the replica count for Deployment
+# Set leaderElection if you want to use more than 1 replica
+# Set affinity.podAntiAffinity rule if you want to schedule onto a node
+# only if that node is in the same zone as at least one already-running descheduler
+replicas: 1
+
+# Specifies whether Leader Election resources should be created
+# Required when running as a Deployment
+# NOTE: Leader election can't be activated if DryRun enabled
+leaderElection: {}
+#  enabled: true
+#  leaseDuration: 15s
+#  renewDeadline: 10s
+#  retryPeriod: 2s
+#  resourceLock: "leases"
+#  resourceName: "descheduler"
+#  resourceNamescape: "kube-system"
+
 cmdOptions:
   v: 3
-  # evict-local-storage-pods:
-  # max-pods-to-evict-per-node: 10
-  # node-selector: "key1=value1,key2=value2"
 
 deschedulerPolicy:
+  # nodeSelector: "key1=value1,key2=value2"
+  # maxNoOfPodsToEvictPerNode: 10
+  # maxNoOfPodsToEvictPerNamespace: 10
+  # ignorePvcPods: true
+  # evictLocalStoragePods: true
   strategies:
     RemoveDuplicates:
       enabled: true
+    RemovePodsHavingTooManyRestarts:
+      enabled: true
+      params:
+        podsHavingTooManyRestarts:
+          podRestartThreshold: 100
+          includingInitContainers: true
     RemovePodsViolatingNodeTaints:
       enabled: true
     RemovePodsViolatingNodeAffinity:
@@ -52,6 +84,10 @@ deschedulerPolicy:
         - requiredDuringSchedulingIgnoredDuringExecution
     RemovePodsViolatingInterPodAntiAffinity:
       enabled: true
+    RemovePodsViolatingTopologySpreadConstraint:
+      enabled: true
+      params:
+        includeSoftConstraints: false
     LowNodeUtilization:
       enabled: true
       params:
@@ -70,6 +106,25 @@ priorityClassName: system-cluster-critical
 nodeSelector: {}
 #  foo: bar
 
+affinity: {}
+# nodeAffinity:
+#   requiredDuringSchedulingIgnoredDuringExecution:
+#     nodeSelectorTerms:
+#     - matchExpressions:
+#       - key: kubernetes.io/e2e-az-name
+#         operator: In
+#         values:
+#         - e2e-az1
+#         - e2e-az2
+#  podAntiAffinity:
+#    requiredDuringSchedulingIgnoredDuringExecution:
+#      - labelSelector:
+#          matchExpressions:
+#            - key: app.kubernetes.io/name
+#              operator: In
+#              values:
+#                - descheduler
+#        topologyKey: "kubernetes.io/hostname"
 tolerations: []
 # - key: 'management'
 #   operator: 'Equal'
@@ -80,13 +135,47 @@ rbac:
   # Specifies whether RBAC resources should be created
   create: true
 
-podSecurityPolicy:
-  # Specifies whether PodSecurityPolicy should be created.
-  create: true
-
 serviceAccount:
   # Specifies whether a ServiceAccount should be created
   create: true
   # The name of the ServiceAccount to use.
   # If not set and create is true, a name is generated using the fullname template
   name:
+  # Specifies custom annotations for the serviceAccount
+  annotations: {}
+
+podAnnotations: {}
+
+podLabels: {}
+
+livenessProbe:
+  failureThreshold: 3
+  httpGet:
+    path: /healthz
+    port: 10258
+    scheme: HTTPS
+  initialDelaySeconds: 3
+  periodSeconds: 10
+
+service:
+  enabled: false
+
+serviceMonitor:
+  enabled: false
+  # The namespace where Prometheus expects to find service monitors.
+  # namespace: ""
+  interval: ""
+  # honorLabels: true
+  insecureSkipVerify: true
+  serverName: null
+  metricRelabelings: []
+    # - action: keep
+    #   regex: 'descheduler_(build_info|pods_evicted)'
+    #   sourceLabels: [__name__]
+  relabelings: []
+    # - sourceLabels: [__meta_kubernetes_pod_node_name]
+    #   separator: ;
+    #   regex: ^(.*)$
+    #   targetLabel: nodename
+    #   replacement: $1
+    #   action: replace

cloudbuild.yaml

@@ -1,13 +1,13 @@
 # See https://cloud.google.com/cloud-build/docs/build-config
 
 # this must be specified in seconds. If omitted, defaults to 600s (10 mins)
-timeout: 1200s
+timeout: 3600s
 # this prevents errors if you don't use both _GIT_TAG and _PULL_BASE_REF,
 # or any new substitutions added in the future.
 options:
   substitution_option: ALLOW_LOOSE
 steps:
-  - name: 'gcr.io/k8s-testimages/gcb-docker-gcloud:v20190906-745fed4'
+  - name: 'gcr.io/k8s-staging-test-infra/gcb-docker-gcloud:v20211118-2f2d816b90'
     entrypoint: make
     env:
     - DOCKER_CLI_EXPERIMENTAL=enabled

cmd/descheduler/app/options/options.go

@@ -18,13 +18,14 @@ limitations under the License.
 package options
 
 import (
-	"github.com/spf13/pflag"
+	"time"
 
-	utilerrors "k8s.io/apimachinery/pkg/util/errors"
+	"github.com/spf13/pflag"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	apiserveroptions "k8s.io/apiserver/pkg/server/options"
 	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/component-base/logs"
-
+	componentbaseconfig "k8s.io/component-base/config"
+	componentbaseoptions "k8s.io/component-base/config/options"
 	"sigs.k8s.io/descheduler/pkg/apis/componentconfig"
 	"sigs.k8s.io/descheduler/pkg/apis/componentconfig/v1alpha1"
 	deschedulerscheme "sigs.k8s.io/descheduler/pkg/descheduler/scheme"
@@ -39,7 +40,7 @@ type DeschedulerServer struct {
 	componentconfig.DeschedulerConfiguration
 
 	Client         clientset.Interface
-	Logs           *logs.Options
+	EventClient    clientset.Interface
 	SecureServing  *apiserveroptions.SecureServingOptionsWithLoopback
 	DisableMetrics bool
 }
@@ -56,20 +57,22 @@ func NewDeschedulerServer() (*DeschedulerServer, error) {
 
 	return &DeschedulerServer{
 		DeschedulerConfiguration: *cfg,
-		Logs:                     logs.NewOptions(),
 		SecureServing:            secureServing,
 	}, nil
 }
 
-// Validation checks for DeschedulerServer.
-func (s *DeschedulerServer) Validate() error {
-	var errs []error
-	errs = append(errs, s.Logs.Validate()...)
-	return utilerrors.NewAggregate(errs)
-}
-
 func newDefaultComponentConfig() (*componentconfig.DeschedulerConfiguration, error) {
-	versionedCfg := v1alpha1.DeschedulerConfiguration{}
+	versionedCfg := v1alpha1.DeschedulerConfiguration{
+		LeaderElection: componentbaseconfig.LeaderElectionConfiguration{
+			LeaderElect:       false,
+			LeaseDuration:     metav1.Duration{Duration: 137 * time.Second},
+			RenewDeadline:     metav1.Duration{Duration: 107 * time.Second},
+			RetryPeriod:       metav1.Duration{Duration: 26 * time.Second},
+			ResourceLock:      "leases",
+			ResourceName:      "descheduler",
+			ResourceNamespace: "kube-system",
+		},
+	}
 	deschedulerscheme.Scheme.Default(&versionedCfg)
 	cfg := componentconfig.DeschedulerConfiguration{}
 	if err := deschedulerscheme.Scheme.Convert(&versionedCfg, &cfg, nil); err != nil {
@@ -80,18 +83,14 @@ func newDefaultComponentConfig() (*componentconfig.DeschedulerConfiguration, err
 
 // AddFlags adds flags for a specific SchedulerServer to the specified FlagSet
 func (rs *DeschedulerServer) AddFlags(fs *pflag.FlagSet) {
-	fs.StringVar(&rs.Logging.Format, "logging-format", "text", `Sets the log format. Permitted formats: "text", "json". Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log-dir, --log-file, --log-file-max-size, --logtostderr, --skip-headers, --skip-log-headers, --stderrthreshold, --log-flush-frequency.\nNon-default choices are currently alpha and subject to change without warning.`)
+	fs.StringVar(&rs.Logging.Format, "logging-format", "text", `Sets the log format. Permitted formats: "text", "json". Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log_dir, --log_file, --log_file_max_size, --logtostderr, --skip-headers, --skip-log-headers, --stderrthreshold, --log-flush-frequency.\nNon-default choices are currently alpha and subject to change without warning.`)
 	fs.DurationVar(&rs.DeschedulingInterval, "descheduling-interval", rs.DeschedulingInterval, "Time interval between two consecutive descheduler executions. Setting this value instructs the descheduler to run in a continuous loop at the interval specified.")
 	fs.StringVar(&rs.KubeconfigFile, "kubeconfig", rs.KubeconfigFile, "File with  kube configuration.")
 	fs.StringVar(&rs.PolicyConfigFile, "policy-config-file", rs.PolicyConfigFile, "File with descheduler policy configuration.")
 	fs.BoolVar(&rs.DryRun, "dry-run", rs.DryRun, "execute descheduler in dry run mode.")
-	// node-selector query causes descheduler to run only on nodes that matches the node labels in the query
-	fs.StringVar(&rs.NodeSelector, "node-selector", rs.NodeSelector, "DEPRECATED: selector (label query) to filter on, supports '=', '==', and '!='.(e.g. -l key1=value1,key2=value2)")
-	// max-no-pods-to-evict limits the maximum number of pods to be evicted per node by descheduler.
-	fs.IntVar(&rs.MaxNoOfPodsToEvictPerNode, "max-pods-to-evict-per-node", rs.MaxNoOfPodsToEvictPerNode, "DEPRECATED: limits the maximum number of pods to be evicted per node by descheduler")
-	// evict-local-storage-pods allows eviction of pods that are using local storage. This is false by default.
-	fs.BoolVar(&rs.EvictLocalStoragePods, "evict-local-storage-pods", rs.EvictLocalStoragePods, "DEPRECATED: enables evicting pods using local storage by descheduler")
 	fs.BoolVar(&rs.DisableMetrics, "disable-metrics", rs.DisableMetrics, "Disables metrics. The metrics are by default served through https://localhost:10258/metrics. Secure address, resp. port can be changed through --bind-address, resp. --secure-port flags.")
 
+	componentbaseoptions.BindLeaderElectionFlags(&rs.LeaderElection, fs)
+
 	rs.SecureServing.AddFlags(fs)
 }

cmd/descheduler/app/server.go

@@ -19,8 +19,12 @@ package app
 
 import (
 	"context"
-	"flag"
 	"io"
+	"os"
+	"os/signal"
+	"syscall"
+
+	"k8s.io/apiserver/pkg/server/healthz"
 
 	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"sigs.k8s.io/descheduler/pkg/descheduler"
@@ -30,7 +34,9 @@ import (
 	apiserver "k8s.io/apiserver/pkg/server"
 	"k8s.io/apiserver/pkg/server/mux"
 	restclient "k8s.io/client-go/rest"
-	aflag "k8s.io/component-base/cli/flag"
+	registry "k8s.io/component-base/logs/api/v1"
+	jsonLog "k8s.io/component-base/logs/json"
+	_ "k8s.io/component-base/logs/json/register"
 	"k8s.io/component-base/metrics/legacyregistry"
 	"k8s.io/klog/v2"
 )
@@ -38,7 +44,6 @@ import (
 // NewDeschedulerCommand creates a *cobra.Command object with default parameters
 func NewDeschedulerCommand(out io.Writer) *cobra.Command {
 	s, err := options.NewDeschedulerServer()
-
 	if err != nil {
 		klog.ErrorS(err, "unable to initialize server")
 	}
@@ -48,8 +53,7 @@ func NewDeschedulerCommand(out io.Writer) *cobra.Command {
 		Short: "descheduler",
 		Long:  `The descheduler evicts pods which may be bound to less desired nodes`,
 		Run: func(cmd *cobra.Command, args []string) {
-			s.Logs.Config.Format = s.Logging.Format
-			s.Logs.Apply()
+			// s.Logs.Config.Format = s.Logging.Format
 
 			// LoopbackClientConfig is a config for a privileged loopback connection
 			var LoopbackClientConfig *restclient.Config
@@ -58,37 +62,59 @@ func NewDeschedulerCommand(out io.Writer) *cobra.Command {
 				klog.ErrorS(err, "failed to apply secure server configuration")
 				return
 			}
+			var factory registry.LogFormatFactory
 
-			if err := s.Validate(); err != nil {
-				klog.ErrorS(err, "failed to validate server configuration")
+			if s.Logging.Format == "json" {
+				factory = jsonLog.Factory{}
+			}
+
+			if factory == nil {
+				klog.ClearLogger()
+			} else {
+				log, logrFlush := factory.Create(registry.LoggingConfiguration{
+					Format:    s.Logging.Format,
+					Verbosity: s.Logging.Verbosity,
+				})
+				defer logrFlush()
+				klog.SetLogger(log)
+			}
+
+			ctx, done := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
+
+			pathRecorderMux := mux.NewPathRecorderMux("descheduler")
+			if !s.DisableMetrics {
+				pathRecorderMux.Handle("/metrics", legacyregistry.HandlerWithReset())
+			}
+
+			healthz.InstallHandler(pathRecorderMux, healthz.NamedCheck("Descheduler", healthz.PingHealthz.Check))
+
+			stoppedCh, _, err := SecureServing.Serve(pathRecorderMux, 0, ctx.Done())
+			if err != nil {
+				klog.Fatalf("failed to start secure server: %v", err)
 				return
 			}
 
-			if !s.DisableMetrics {
-				ctx := context.TODO()
-				pathRecorderMux := mux.NewPathRecorderMux("descheduler")
-				pathRecorderMux.Handle("/metrics", legacyregistry.HandlerWithReset())
-
-				if _, err := SecureServing.Serve(pathRecorderMux, 0, ctx.Done()); err != nil {
-					klog.Fatalf("failed to start secure server: %v", err)
-					return
-				}
-			}
-
-			err := Run(s)
+			err = Run(ctx, s)
 			if err != nil {
 				klog.ErrorS(err, "descheduler server")
 			}
+
+			done()
+			// wait for metrics server to close
+			<-stoppedCh
 		},
 	}
 	cmd.SetOut(out)
 	flags := cmd.Flags()
-	flags.SetNormalizeFunc(aflag.WordSepNormalizeFunc)
-	flags.AddGoFlagSet(flag.CommandLine)
 	s.AddFlags(flags)
 	return cmd
 }
 
-func Run(rs *options.DeschedulerServer) error {
-	return descheduler.Run(rs)
+func Run(ctx context.Context, rs *options.DeschedulerServer) error {
+	return descheduler.Run(ctx, rs)
+}
+
+func SetupLogs() {
+	klog.SetOutput(os.Stdout)
+	klog.InitFlags(nil)
 }

cmd/descheduler/app/version.go

@@ -24,7 +24,7 @@ import (
 )
 
 func NewVersionCommand() *cobra.Command {
-	var versionCmd = &cobra.Command{
+	versionCmd := &cobra.Command{
 		Use:   "version",
 		Short: "Version of descheduler",
 		Long:  `Prints the version of descheduler.`,

cmd/descheduler/descheduler.go

@@ -17,22 +17,23 @@ limitations under the License.
 package main
 
 import (
-	"fmt"
-	"k8s.io/component-base/logs"
 	"os"
+
+	"k8s.io/component-base/cli"
 	"sigs.k8s.io/descheduler/cmd/descheduler/app"
+	"sigs.k8s.io/descheduler/pkg/descheduler"
 )
 
+func init() {
+	app.SetupLogs()
+	descheduler.SetupPlugins()
+}
+
 func main() {
 	out := os.Stdout
 	cmd := app.NewDeschedulerCommand(out)
 	cmd.AddCommand(app.NewVersionCommand())
 
-	logs.InitLogs()
-	defer logs.FlushLogs()
-
-	if err := cmd.Execute(); err != nil {
-		fmt.Println(err)
-		os.Exit(1)
-	}
+	code := cli.Run(cmd)
+	os.Exit(code)
 }

docs/contributor-guide.md

@@ -31,7 +31,7 @@ View all CLI options.
 ## Run Tests
 ```
 GOOS=linux make dev-image
-kind create cluster --config hack/kind_config.yaml
+make kind-multi-node
 kind load docker-image <image name>
 kind get kubeconfig > /tmp/admin.conf
 export KUBECONFIG=/tmp/admin.conf
@@ -39,17 +39,38 @@ make test-unit
 make test-e2e
 ```
 
-## Run Helm Tests
-Run the helm test for a particular descheduler release by setting below variables,
-```
-HELM_IMAGE_REPO="descheduler"
-HELM_IMAGE_TAG="helm-test"
-HELM_CHART_LOCATION="./charts/descheduler"
-```
-The helm tests runs as part of descheduler CI. But, to run it manually from the descheduler root,
+## Format Code
+
+After making changes in the code base, ensure that the code is formatted correctly:
 
 ```
-make test-helm
+make fmt
+```
+
+## Build Helm Package locally
+
+If you made some changes in the chart, and just want to check if templating is ok, or if the chart is buildable, you can run this command to have a package built from the `./charts` directory.
+
+```
+make build-helm
+```
+
+## Lint Helm Chart locally
+
+To check linting of your changes in the helm chart locally you can run:
+
+```
+make lint-chart
+```
+
+## Test helm changes locally with kind and ct
+
+You will need kind and docker (or equivalent) installed. We can use ct public image to avoid installing ct and all its dependencies.
+
+
+```
+make kind-multi-node
+make ct-helm
 ```
 
 ### Miscellaneous

docs/proposals.md

@@ -0,0 +1,16 @@
+# Proposals
+This document walk you through about all the enhancements proposals for descheduler.
+
+## Descheduler v1alpha2 Design Proposal
+```yaml
+title: Descheduler v1alpha2 Design Proposal
+authors:
+- "@damemi"
+link: 
+- https://docs.google.com/document/d/1S1JCh-0F-QCJvBBG-kbmXiHAJFF8doArhDIAKbOj93I/edit#heading=h.imbp1ctnc8lx
+- https://github.com/kubernetes-sigs/descheduler/issues/679
+owning-sig: sig-scheduling
+creation-date: 2021-05-01
+status: implementable
+```
+

docs/release-guide.md

@@ -1,36 +1,82 @@
 # Release Guide
 
-## Container Image
+The process for publishing each Descheduler release includes a mixture of manual and automatic steps. Over 
+time, it would be good to automate as much of this process as possible. However, due to current limitations there 
+is care that must be taken to perform each manual step precisely so that the automated steps execute properly.
 
-### Semi-automatic
+## Pre-release Code Changes
 
-1. Make sure your repo is clean by git's standards
-2. Create a release branch `git checkout -b release-1.18` (not required for patch releases)
-3. Push the release branch to the descheuler repo and ensure branch protection is enabled (not required for patch releases)
-4. Tag the repository from the `master` branch (from the `release-1.18` branch for a patch release) and push the tag `VERSION=v0.18.0 git tag -m $VERSION $VERSION; git push origin $VERSION`
-5. Publish a draft release using the tag you just created
-6. Perform the [image promotion process](https://github.com/kubernetes/k8s.io/tree/main/k8s.gcr.io#image-promoter)
-7. Publish release
-8. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
+Before publishing each release, the following code updates must be made:
 
-### Manual
+- [ ] (Optional, but recommended) Bump `k8s.io` dependencies to the `-rc` tags. These tags are usually published around upstream code freeze. [Example](https://github.com/kubernetes-sigs/descheduler/pull/539)
+- [ ] Bump `k8s.io` dependencies to GA tags once they are published (following the upstream release). [Example](https://github.com/kubernetes-sigs/descheduler/pull/615)
+- [ ] Ensure that Go is updated to the same version as upstream. [Example](https://github.com/kubernetes-sigs/descheduler/pull/801)
+- [ ] Make CI changes in [github.com/kubernetes/test-infra](https://github.com/kubernetes/test-infra) to add the new version's tests (note, this may also include a Go bump). [Example](https://github.com/kubernetes/test-infra/pull/25833)
+- [ ] Update local CI versions for utils (such as golang-ci), kind, and go. [Example - e2e](https://github.com/kubernetes-sigs/descheduler/commit/ac4d576df8831c0c399ee8fff1e85469e90b8c44), [Example - helm](https://github.com/kubernetes-sigs/descheduler/pull/821)
+- [ ] Update version references in docs and Readme. [Example](https://github.com/kubernetes-sigs/descheduler/pull/617)
 
-1. Make sure your repo is clean by git's standards
-2. Create a release branch `git checkout -b release-1.18` (not required for patch releases)
-3. Push the release branch to the descheuler repo and ensure branch protection is enabled (not required for patch releases)
-4. Tag the repository from the `master` branch (from the `release-1.18` branch for a patch release) and push the tag `VERSION=v0.18.0 git tag -m $VERSION $VERSION; git push origin $VERSION`
-5. Checkout the tag you just created and make sure your repo is clean by git's standards `git checkout $VERSION`
-6. Build and push the container image to the staging registry `VERSION=$VERSION make push-all`
-7. Publish a draft release using the tag you just created
-8. Perform the [image promotion process](https://github.com/kubernetes/k8s.io/tree/main/k8s.gcr.io#image-promoter)
-9. Publish release
-10. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
+## Release Process
 
-### Notes
-It's important to create the tag on the master branch after creating the release-* branch so that the [Helm releaser-action](#helm-chart) can work.
-It compares the changes in the action-triggering branch to the latest tag on that branch, so if you tag before creating the new branch there
-will be nothing to compare and it will fail (creating a new release branch usually involves no code changes). For this same reason, you should
-also tag patch releases (on the release-* branch) *after* pushing changes (if those changes involve bumping the Helm chart version).
+When the above pre-release steps are complete and the release is ready to be cut, perform the following steps **in order** 
+(the flowchart below demonstrates these steps):
+
+**Version release**
+1. Create the `git tag` on `master` for the release, eg `v0.24.0`
+2. Merge Helm chart version update to `master` (see [Helm chart](#helm-chart) below). [Example](https://github.com/kubernetes-sigs/descheduler/pull/709)
+3. Perform the [image promotion process](https://github.com/kubernetes/k8s.io/tree/main/k8s.gcr.io#image-promoter). [Example](https://github.com/kubernetes/k8s.io/pull/3344)
+4. Cut release branch from `master`, eg `release-1.24`
+5. Publish release using Github's release process from the git tag you created
+6. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
+
+**Patch release**
+1. Pick relevant code change commits to the matching release branch, eg `release-1.24`
+2. Create the patch tag on the release branch, eg `v0.24.1` on `release-1.24`
+3. Merge Helm chart version update to release branch
+4. Perform the image promotion process for the patch version
+5. Publish release using Github's release process from the git tag you created
+6. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
+
+### Flowchart
+
+![Flowchart for major and patch releases](release-process.png)
+
+### Image promotion process
+
+Every merge to any branch triggers an [image build and push](https://github.com/kubernetes/test-infra/blob/c36b8e5/config/jobs/image-pushing/k8s-staging-descheduler.yaml) to a `gcr.io` repository. 
+These automated image builds are snapshots of the code in place at the time of every PR merge and 
+tagged with the latest git SHA at the time of the build. To create a final release image, the desired 
+auto-built image SHA is added to a [file upstream](https://github.com/kubernetes/k8s.io/blob/e9e971c/k8s.gcr.io/images/k8s-staging-descheduler/images.yaml) which 
+copies that image to a public registry.
+
+Automatic builds can be monitored and re-triggered with the [`post-descheduler-push-images` job](https://prow.k8s.io/?job=post-descheduler-push-images) on prow.k8s.io.
+
+Note that images can also be manually built and pushed using `VERSION=$VERSION make push-all` by [users with access](https://github.com/kubernetes/k8s.io/blob/fbee8f67b70304241e613a672c625ad972998ad7/groups/sig-scheduling/groups.yaml#L33-L43).
+
+## Helm Chart
+We currently use the [chart-releaser-action GitHub Action](https://github.com/helm/chart-releaser-action) to automatically 
+publish [Helm chart releases](https://github.com/kubernetes-sigs/descheduler/blob/022e07c/.github/workflows/release.yaml). 
+This action is triggered when it detects any changes to [`Chart.yaml`](https://github.com/kubernetes-sigs/descheduler/blob/022e07c27853fade6d1304adc0a6ebe02642386c/charts/descheduler/Chart.yaml) on 
+a `release-*` branch.
+
+Helm chart releases are managed by a separate set of git tags that are prefixed with `descheduler-helm-chart-*`. Example git tag name is `descheduler-helm-chart-0.18.0`.
+Released versions of the helm charts are stored in the `gh-pages` branch of this repo.
+
+The major and minor version of the chart matches the descheduler major and minor versions. For example descheduler helm chart version helm-descheduler-chart-0.18.0 corresponds
+to descheduler version v0.18.0. The patch version of the descheduler helm chart and the patcher version of the descheduler will not necessarily match. The patch
+version of the descheduler helm chart is used to version changes specific to the helm chart.
+
+1. Merge all helm chart changes into the master branch before the release is tagged/cut
+   1. Ensure that `appVersion` in file `charts/descheduler/Chart.yaml` matches the descheduler version(no `v` prefix)
+   2. Ensure that `version` in file `charts/descheduler/Chart.yaml` has been incremented. This is the chart version.
+2. Make sure your repo is clean by git's standards
+3. Follow the release-branch or patch release tagging pattern from the above section.
+4. Verify the new helm artifact has been successfully pushed to the `gh-pages` branch
+
+## Notes
+The Helm releaser-action compares the changes in the action-triggering branch to the latest tag on that branch, so if you tag before creating the new branch there
+will be nothing to compare and it will fail. This is why it's necessary to tag, eg, `v0.24.0` *before* making the changes to the 
+Helm chart version, so that there is a new diff for the action to find. (Tagging *after* making the Helm chart changes would 
+also work, but then the code that gets built into the promoted image will be tagged as `descheduler-helm-chart-xxx` rather than `v0.xx.0`).
 
 See [post-descheduler-push-images dashboard](https://testgrid.k8s.io/sig-scheduling#post-descheduler-push-images) for staging registry image build job status.
 
@@ -56,19 +102,3 @@ Pull image from the staging registry.
 ```
 docker pull gcr.io/k8s-staging-descheduler/descheduler:v20200206-0.9.0-94-ge2a23f284
 ```
-
-## Helm Chart
-Helm chart releases are managed by a separate set of git tags that are prefixed with `descheduler-helm-chart-*`. Example git tag name is `descheduler-helm-chart-0.18.0`.
-Released versions of the helm charts are stored in the `gh-pages` branch of this repo. The [chart-releaser-action GitHub Action](https://github.com/helm/chart-releaser-action)
-is setup to build and push the helm charts to the `gh-pages` branch when changes are pushed to a `release-*` branch.
-
-The major and minor version of the chart matches the descheduler major and minor versions. For example descheduler helm chart version helm-descheduler-chart-0.18.0 corresponds
-to descheduler version v0.18.0. The patch version of the descheduler helm chart and the patcher version of the descheduler will not necessarily match. The patch
-version of the descheduler helm chart is used to version changes specific to the helm chart.
-
-1. Merge all helm chart changes into the master branch before the release is tagged/cut
-   1. Ensure that `appVersion` in file `charts/descheduler/Chart.yaml` matches the descheduler version(no `v` prefix)
-   2. Ensure that `version` in file `charts/descheduler/Chart.yaml` has been incremented. This is the chart version.
-2. Make sure your repo is clean by git's standards
-3. Follow the release-branch or patch release tagging pattern from the above section.
-4. Verify the new helm artifact has been successfully pushed to the `gh-pages` branch

docs/user-guide.md

@@ -2,22 +2,27 @@
 
 Starting with descheduler release v0.10.0 container images are available in the official k8s container registry.
 
-Descheduler Version | Container Image                                     | Architectures           |
-------------------- |-----------------------------------------------------|-------------------------|
-v0.22.0             | k8s.gcr.io/descheduler/descheduler:v0.22.0          | AMD64<br>ARM64<br>ARMv7 |
-v0.21.0             | k8s.gcr.io/descheduler/descheduler:v0.21.0          | AMD64<br>ARM64<br>ARMv7 |
-v0.20.0             | k8s.gcr.io/descheduler/descheduler:v0.20.0          | AMD64<br>ARM64          |
-v0.19.0             | k8s.gcr.io/descheduler/descheduler:v0.19.0          | AMD64                   |
-v0.18.0             | k8s.gcr.io/descheduler/descheduler:v0.18.0          | AMD64                   |
-v0.10.0             | k8s.gcr.io/descheduler/descheduler:v0.10.0          | AMD64                   |
+Descheduler Version | Container Image                                 | Architectures           |
+------------------- |-------------------------------------------------|-------------------------|
+v0.25.1             | registry.k8s.io/descheduler/descheduler:v0.25.1 | AMD64<br>ARM64<br>ARMv7 |
+v0.25.0             | registry.k8s.io/descheduler/descheduler:v0.25.0 | AMD64<br>ARM64<br>ARMv7 |
+v0.24.1             | registry.k8s.io/descheduler/descheduler:v0.24.1 | AMD64<br>ARM64<br>ARMv7 |
+v0.24.0             | registry.k8s.io/descheduler/descheduler:v0.24.0 | AMD64<br>ARM64<br>ARMv7 |
+v0.23.1             | registry.k8s.io/descheduler/descheduler:v0.23.1 | AMD64<br>ARM64<br>ARMv7 |
+v0.22.0             | registry.k8s.io/descheduler/descheduler:v0.22.0 | AMD64<br>ARM64<br>ARMv7 |
+v0.21.0             | registry.k8s.io/descheduler/descheduler:v0.21.0 | AMD64<br>ARM64<br>ARMv7 |
+v0.20.0             | registry.k8s.io/descheduler/descheduler:v0.20.0 | AMD64<br>ARM64          |
+v0.19.0             | registry.k8s.io/descheduler/descheduler:v0.19.0 | AMD64                   |
+v0.18.0             | registry.k8s.io/descheduler/descheduler:v0.18.0 | AMD64                   |
+v0.10.0             | registry.k8s.io/descheduler/descheduler:v0.10.0 | AMD64                   |
 
 Note that multi-arch container images cannot be pulled by [kind](https://kind.sigs.k8s.io) from a registry. Therefore
 starting with descheduler release v0.20.0 use the below process to download the official descheduler
 image into a kind cluster.
 ```
 kind create cluster
-docker pull k8s.gcr.io/descheduler/descheduler:v0.20.0
-kind load docker-image k8s.gcr.io/descheduler/descheduler:v0.20.0
+docker pull registry.k8s.io/descheduler/descheduler:v0.20.0
+kind load docker-image registry.k8s.io/descheduler/descheduler:v0.20.0
 ```
 
 ## Policy Configuration Examples
@@ -34,31 +39,52 @@ Usage:
   descheduler [command]
 
 Available Commands:
+  completion  generate the autocompletion script for the specified shell
   help        Help about any command
   version     Version of descheduler
 
 Flags:
-      --add-dir-header                   If true, adds the file directory to the header of the log messages
-      --alsologtostderr                  log to standard error as well as files
-      --descheduling-interval duration   Time interval between two consecutive descheduler executions. Setting this value instructs the descheduler to run in a continuous loop at the interval specified.
-      --dry-run                          execute descheduler in dry run mode.
-      --evict-local-storage-pods         DEPRECATED: enables evicting pods using local storage by descheduler
-  -h, --help                             help for descheduler
-      --kubeconfig string                File with  kube configuration.
-      --log-backtrace-at traceLocation   when logging hits line file:N, emit a stack trace (default :0)
-      --log-dir string                   If non-empty, write log files in this directory
-      --log-file string                  If non-empty, use this log file
-      --log-file-max-size uint           Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited. (default 1800)
-      --log-flush-frequency duration     Maximum number of seconds between log flushes (default 5s)
-      --logtostderr                      log to standard error instead of files (default true)
-      --max-pods-to-evict-per-node int   DEPRECATED: limits the maximum number of pods to be evicted per node by descheduler
-      --node-selector string             DEPRECATED: selector (label query) to filter on, supports '=', '==', and '!='.(e.g. -l key1=value1,key2=value2)
-      --policy-config-file string        File with descheduler policy configuration.
-      --skip-headers                     If true, avoid header prefixes in the log messages
-      --skip-log-headers                 If true, avoid headers when opening log files
-      --stderrthreshold severity         logs at or above this threshold go to stderr (default 2)
-  -v, --v Level                          number for the log level verbosity
-      --vmodule moduleSpec               comma-separated list of pattern=N settings for file-filtered logging
+      --add-dir-header                           If true, adds the file directory to the header of the log messages (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --alsologtostderr                          log to standard error as well as files (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --bind-address ip                          The IP address on which to listen for the --secure-port port. The associated interface(s) must be reachable by the rest of the cluster, and by CLI/web clients. If blank or an unspecified address (0.0.0.0 or ::), all interfaces will be used. (default 0.0.0.0)
+      --cert-dir string                          The directory where the TLS certs are located. If --tls-cert-file and --tls-private-key-file are provided, this flag will be ignored. (default "apiserver.local.config/certificates")
+      --descheduling-interval duration           Time interval between two consecutive descheduler executions. Setting this value instructs the descheduler to run in a continuous loop at the interval specified.
+      --disable-metrics                          Disables metrics. The metrics are by default served through https://localhost:10258/metrics. Secure address, resp. port can be changed through --bind-address, resp. --secure-port flags.
+      --dry-run                                  execute descheduler in dry run mode.
+  -h, --help                                     help for descheduler
+      --http2-max-streams-per-connection int     The limit that the server gives to clients for the maximum number of streams in an HTTP/2 connection. Zero means to use golang's default.
+      --kubeconfig string                        File with  kube configuration.
+      --leader-elect                             Start a leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability.
+      --leader-elect-lease-duration duration     The duration that non-leader candidates will wait after observing a leadership renewal until attempting to acquire leadership of a led but unrenewed leader slot. This is effectively the maximum duration that a leader can be stopped before it is replaced by another candidate. This is only applicable if leader election is enabled. (default 15s)
+      --leader-elect-renew-deadline duration     The interval between attempts by the acting master to renew a leadership slot before it stops leading. This must be less than or equal to the lease duration. This is only applicable if leader election is enabled. (default 10s)
+      --leader-elect-resource-lock string        The type of resource object that is used for locking during leader election. Supported options are 'endpoints', 'configmaps', 'leases', 'endpointsleases' and 'configmapsleases'. (default "leases")
+      --leader-elect-resource-name string        The name of resource object that is used for locking during leader election. (default "descheduler")
+      --leader-elect-resource-namespace string   The namespace of resource object that is used for locking during leader election. (default "kube-system")
+      --leader-elect-retry-period duration       The duration the clients should wait between attempting acquisition and renewal of a leadership. This is only applicable if leader election is enabled. (default 2s)
+      --log-backtrace-at traceLocation           when logging hits line file:N, emit a stack trace (default :0) (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --log_dir string                           If non-empty, write log files in this directory (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --log_file string                          If non-empty, use this log file (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --log_file_max_size uint                   Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited. (default 1800) (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --log-flush-frequency duration             Maximum number of seconds between log flushes (default 5s)
+      --logging-format string                    Sets the log format. Permitted formats: "text", "json". Non-default formats don't honor these flags: --add-dir-header, --alsologtostderr, --log-backtrace-at, --log_dir, --log_file, --log_file_max_size, --logtostderr, --skip-headers, --skip-log-headers, --stderrthreshold, --log-flush-frequency.\nNon-default choices are currently alpha and subject to change without warning. (default "text")
+      --logtostderr                              log to standard error instead of files (default true) (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --one-output                               If true, only write logs to their native severity level (vs also writing to each lower severity level) (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --permit-address-sharing                   If true, SO_REUSEADDR will be used when binding the port. This allows binding to wildcard IPs like 0.0.0.0 and specific IPs in parallel, and it avoids waiting for the kernel to release sockets in TIME_WAIT state. [default=false]
+      --permit-port-sharing                      If true, SO_REUSEPORT will be used when binding the port, which allows more than one instance to bind on the same address and port. [default=false]
+      --policy-config-file string                File with descheduler policy configuration.
+      --secure-port int                          The port on which to serve HTTPS with authentication and authorization. If 0, don't serve HTTPS at all. (default 10258)
+      --skip-headers                             If true, avoid header prefixes in the log messages (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --skip-log-headers                         If true, avoid headers when opening log files (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --stderrthreshold severity                 logs at or above this threshold go to stderr (default 2) (DEPRECATED: will be removed in a future release, see https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/2845-deprecate-klog-specific-flags-in-k8s-components)
+      --tls-cert-file string                     File containing the default x509 Certificate for HTTPS. (CA cert, if any, concatenated after server cert). If HTTPS serving is enabled, and --tls-cert-file and --tls-private-key-file are not provided, a self-signed certificate and key are generated for the public address and saved to the directory specified by --cert-dir.
+      --tls-cipher-suites strings                Comma-separated list of cipher suites for the server. If omitted, the default Go cipher suites will be used. 
+                                                 Preferred values: TLS_AES_128_GCM_SHA256, TLS_AES_256_GCM_SHA384, TLS_CHACHA20_POLY1305_SHA256, TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305, TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256, TLS_RSA_WITH_AES_128_CBC_SHA, TLS_RSA_WITH_AES_128_GCM_SHA256, TLS_RSA_WITH_AES_256_CBC_SHA, TLS_RSA_WITH_AES_256_GCM_SHA384. 
+                                                 Insecure values: TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_ECDSA_WITH_RC4_128_SHA, TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_RSA_WITH_RC4_128_SHA, TLS_RSA_WITH_3DES_EDE_CBC_SHA, TLS_RSA_WITH_AES_128_CBC_SHA256, TLS_RSA_WITH_RC4_128_SHA.
+      --tls-min-version string                   Minimum TLS version supported. Possible values: VersionTLS10, VersionTLS11, VersionTLS12, VersionTLS13
+      --tls-private-key-file string              File containing the default x509 private key matching --tls-cert-file.
+      --tls-sni-cert-key namedCertKey            A pair of x509 certificate and private key file paths, optionally suffixed with a list of domain patterns which are fully qualified domain names, possibly with prefixed wildcard segments. The domain patterns also allow IP addresses, but IPs should only be used if the apiserver has visibility to the IP address requested by a client. If no domain patterns are provided, the names of the certificate are extracted. Non-wildcard matches trump over wildcard matches, explicit domain patterns trump over extracted names. For multiple key/certificate pairs, use the --tls-sni-cert-key multiple times. Examples: "example.crt,example.key" or "foo.crt,foo.key:*.foo.com,foo.com". (default [])
+  -v, --v Level                                  number for the log level verbosity
+      --vmodule moduleSpec                       comma-separated list of pattern=N settings for file-filtered logging
 
 Use "descheduler [command] --help" for more information about a command.
 ```
@@ -89,7 +115,8 @@ strategies:
   "PodLifeTime":
     enabled: true
     params:
-      maxPodLifeTimeSeconds: 604800 # pods run for a maximum of 7 days
+      podLifeTime:
+        maxPodLifeTimeSeconds: 604800 # pods run for a maximum of 7 days
 ```
 
 ### Balance Cluster By Node Memory Utilization
@@ -117,7 +144,7 @@ strategies:
 
 #### Balance low utilization nodes
 Using `HighNodeUtilization`, descheduler will rebalance the cluster based on memory by evicting pods
-from nodes with memory utilization lower than 20%. This should be used along with scheduler strategy `MostRequestedPriority`.
+from nodes with memory utilization lower than 20%. This should be use `NodeResourcesFit` with the `MostAllocated` scoring strategy based on these [doc](https://kubernetes.io/docs/reference/scheduling/config/#scheduling-plugins).
 The evicted pods will be compacted into minimal set of nodes.
 
 ```
@@ -133,10 +160,19 @@ strategies:
 ```
 
 ### Autoheal Node Problems
+
 Descheduler's `RemovePodsViolatingNodeTaints` strategy can be combined with
 [Node Problem Detector](https://github.com/kubernetes/node-problem-detector/) and
 [Cluster Autoscaler](https://github.com/kubernetes/autoscaler/tree/master/cluster-autoscaler) to automatically remove
-Nodes which have problems. Node Problem Detector can detect specific Node problems and taint any Nodes which have those
-problems. The Descheduler will then deschedule workloads from those Nodes. Finally, if the descheduled Node's resource
+Nodes which have problems. Node Problem Detector can detect specific Node problems and report them to the API server.
+There is a feature called TaintNodeByCondition of the node controller that takes some conditions and turns them into taints. Currently, this only works for the default node conditions: PIDPressure, MemoryPressure, DiskPressure, Ready, and some cloud provider specific conditions.
+The Descheduler will then deschedule workloads from those Nodes. Finally, if the descheduled Node's resource
 allocation falls below the Cluster Autoscaler's scale down threshold, the Node will become a scale down candidate
 and can be removed by Cluster Autoscaler. These three components form an autohealing cycle for Node problems.
+
+---
+**NOTE**
+
+Once [kubernetes/node-problem-detector#565](https://github.com/kubernetes/node-problem-detector/pull/565) is available in NPD, we need to update this section.
+
+---

examples/failed-pods.yaml

@@ -11,4 +11,4 @@ strategies:
         includingInitContainers: true
         excludeOwnerKinds:
         - "Job"
-        minPodLifeTimeSeconds: 3600 # 1 hour
+        minPodLifetimeSeconds: 3600 # 1 hour

examples/pod-life-time.yml

@@ -6,3 +6,6 @@ strategies:
     params:
       podLifeTime:
         maxPodLifeTimeSeconds: 604800 # 7 days
+        states:
+        - "Pending"
+        - "PodInitializing"

examples/topology-spread-constraint.yaml

@@ -4,4 +4,5 @@ strategies:
   "RemovePodsViolatingTopologySpreadConstraint":
      enabled: true
      params:
+       nodeFit: true
        includeSoftConstraints: true # Include 'ScheduleAnyways' constraints

go.mod

@@ -1,19 +1,110 @@
 module sigs.k8s.io/descheduler
 
-go 1.16
+go 1.19
 
 require (
 	github.com/client9/misspell v0.3.4
-	github.com/spf13/cobra v1.1.3
+	github.com/google/go-cmp v0.5.9
+	github.com/spf13/cobra v1.6.0
 	github.com/spf13/pflag v1.0.5
-	k8s.io/api v0.22.0
-	k8s.io/apimachinery v0.22.0
-	k8s.io/apiserver v0.22.0
-	k8s.io/client-go v0.22.0
-	k8s.io/code-generator v0.22.0
-	k8s.io/component-base v0.22.0
-	k8s.io/component-helpers v0.22.0
-	k8s.io/klog/v2 v2.9.0
-	k8s.io/kubectl v0.20.5
+	k8s.io/api v0.26.0
+	k8s.io/apimachinery v0.26.0
+	k8s.io/apiserver v0.26.0
+	k8s.io/client-go v0.26.0
+	k8s.io/code-generator v0.26.0
+	k8s.io/component-base v0.26.0
+	k8s.io/component-helpers v0.26.0
+	k8s.io/klog/v2 v2.80.1
+	k8s.io/utils v0.0.0-20221128185143-99ec85e7a448
 	sigs.k8s.io/mdtoc v1.0.1
 )
+
+require (
+	github.com/BurntSushi/toml v0.3.1 // indirect
+	github.com/NYTimes/gziphandler v1.1.1 // indirect
+	github.com/antlr/antlr4/runtime/Go/antlr v1.4.10 // indirect
+	github.com/beorn7/perks v1.0.1 // indirect
+	github.com/blang/semver/v4 v4.0.0 // indirect
+	github.com/cenkalti/backoff/v4 v4.1.3 // indirect
+	github.com/cespare/xxhash/v2 v2.1.2 // indirect
+	github.com/coreos/go-semver v0.3.0 // indirect
+	github.com/coreos/go-systemd/v22 v22.3.2 // indirect
+	github.com/davecgh/go-spew v1.1.1 // indirect
+	github.com/emicklei/go-restful/v3 v3.9.0 // indirect
+	github.com/evanphx/json-patch v4.12.0+incompatible // indirect
+	github.com/felixge/httpsnoop v1.0.3 // indirect
+	github.com/fsnotify/fsnotify v1.6.0 // indirect
+	github.com/go-logr/logr v1.2.3 // indirect
+	github.com/go-logr/stdr v1.2.2 // indirect
+	github.com/go-logr/zapr v1.2.3 // indirect
+	github.com/go-openapi/jsonpointer v0.19.5 // indirect
+	github.com/go-openapi/jsonreference v0.20.0 // indirect
+	github.com/go-openapi/swag v0.19.14 // indirect
+	github.com/gogo/protobuf v1.3.2 // indirect
+	github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da // indirect
+	github.com/golang/protobuf v1.5.2 // indirect
+	github.com/gomarkdown/markdown v0.0.0-20200824053859-8c8b3816f167 // indirect
+	github.com/google/cel-go v0.12.5 // indirect
+	github.com/google/gnostic v0.5.7-v3refs // indirect
+	github.com/google/gofuzz v1.1.0 // indirect
+	github.com/google/uuid v1.1.2 // indirect
+	github.com/grpc-ecosystem/go-grpc-prometheus v1.2.0 // indirect
+	github.com/grpc-ecosystem/grpc-gateway/v2 v2.7.0 // indirect
+	github.com/imdario/mergo v0.3.6 // indirect
+	github.com/inconshreveable/mousetrap v1.0.1 // indirect
+	github.com/josharian/intern v1.0.0 // indirect
+	github.com/json-iterator/go v1.1.12 // indirect
+	github.com/mailru/easyjson v0.7.6 // indirect
+	github.com/matttproud/golang_protobuf_extensions v1.0.2 // indirect
+	github.com/mmarkdown/mmark v2.0.40+incompatible // indirect
+	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
+	github.com/modern-go/reflect2 v1.0.2 // indirect
+	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
+	github.com/pkg/errors v0.9.1 // indirect
+	github.com/prometheus/client_golang v1.14.0 // indirect
+	github.com/prometheus/client_model v0.3.0 // indirect
+	github.com/prometheus/common v0.37.0 // indirect
+	github.com/prometheus/procfs v0.8.0 // indirect
+	github.com/stoewer/go-strcase v1.2.0 // indirect
+	go.etcd.io/etcd/api/v3 v3.5.5 // indirect
+	go.etcd.io/etcd/client/pkg/v3 v3.5.5 // indirect
+	go.etcd.io/etcd/client/v3 v3.5.5 // indirect
+	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.35.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.35.0 // indirect
+	go.opentelemetry.io/otel v1.10.0 // indirect
+	go.opentelemetry.io/otel/exporters/otlp/internal/retry v1.10.0 // indirect
+	go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.10.0 // indirect
+	go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracegrpc v1.10.0 // indirect
+	go.opentelemetry.io/otel/metric v0.31.0 // indirect
+	go.opentelemetry.io/otel/sdk v1.10.0 // indirect
+	go.opentelemetry.io/otel/trace v1.10.0 // indirect
+	go.opentelemetry.io/proto/otlp v0.19.0 // indirect
+	go.uber.org/atomic v1.7.0 // indirect
+	go.uber.org/multierr v1.6.0 // indirect
+	go.uber.org/zap v1.19.0 // indirect
+	golang.org/x/crypto v0.1.0 // indirect
+	golang.org/x/mod v0.6.0 // indirect
+	golang.org/x/net v0.3.1-0.20221206200815-1e63c2f08a10 // indirect
+	golang.org/x/oauth2 v0.0.0-20220223155221-ee480838109b // indirect
+	golang.org/x/sync v0.0.0-20220722155255-886fb9371eb4 // indirect
+	golang.org/x/sys v0.3.0 // indirect
+	golang.org/x/term v0.3.0 // indirect
+	golang.org/x/text v0.5.0 // indirect
+	golang.org/x/time v0.0.0-20220210224613-90d013bbcef8 // indirect
+	golang.org/x/tools v0.2.0 // indirect
+	google.golang.org/appengine v1.6.7 // indirect
+	google.golang.org/genproto v0.0.0-20220502173005-c8bf987b8c21 // indirect
+	google.golang.org/grpc v1.49.0 // indirect
+	google.golang.org/protobuf v1.28.1 // indirect
+	gopkg.in/inf.v0 v0.9.1 // indirect
+	gopkg.in/natefinch/lumberjack.v2 v2.0.0 // indirect
+	gopkg.in/yaml.v2 v2.4.0 // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
+	k8s.io/gengo v0.0.0-20220902162205-c0856e24416d // indirect
+	k8s.io/kms v0.26.0 // indirect
+	k8s.io/kube-openapi v0.0.0-20221012153701-172d655c2280 // indirect
+	sigs.k8s.io/apiserver-network-proxy/konnectivity-client v0.0.33 // indirect
+	sigs.k8s.io/json v0.0.0-20220713155537-f223a00ba0e2 // indirect
+	sigs.k8s.io/structured-merge-diff/v4 v4.2.3 // indirect
+	sigs.k8s.io/yaml v1.3.0 // indirect
+)

hack/lib/generator-help.sh

@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+
+# Utility command based on 'find' command. The pipeline is as following:
+#   1. find all the go files; (exclude specific path: vendor etc)
+#   2. find all the files containing specific tags in contents;
+#   3. extract related dirs;
+#   4. remove duplicated paths;
+#   5. merge all dirs in array with delimiter ,;
+#
+# Example:
+#   find_dirs_containing_comment_tags("+k8s:")
+# Return:
+#   sigs.k8s.io/descheduler/a,sigs.k8s.io/descheduler/b,sigs.k8s.io/descheduler/c
+function find_dirs_containing_comment_tags() {
+   array=()
+   while IFS='' read -r line; do array+=("$line"); done < <( \
+     find . -type f -name \*.go -not -path "./vendor/*" -not -path "./_tmp/*" -print0  \
+     | xargs -0 grep --color=never -l "$@" \
+     | xargs -n1 dirname \
+     | LC_ALL=C sort -u \
+     )
+
+   IFS=",";
+   printf '%s' "${array[*]}";
+}

hack/tools.go

@@ -1,3 +1,4 @@
+//go:build tools
 // +build tools
 
 /*

hack/update-generated-conversions.sh

@@ -1,9 +1,10 @@
 #!/bin/bash
 source "$(dirname "${BASH_SOURCE}")/lib/init.sh"
+source "$(dirname "${BASH_SOURCE}")/lib/generator-help.sh"
 
 go build -o "${OS_OUTPUT_BINPATH}/conversion-gen" "k8s.io/code-generator/cmd/conversion-gen"
 
 ${OS_OUTPUT_BINPATH}/conversion-gen \
 		--go-header-file "hack/boilerplate/boilerplate.go.txt" \
-		--input-dirs "${PRJ_PREFIX}/pkg/apis/componentconfig/v1alpha1,${PRJ_PREFIX}/pkg/api/v1alpha1" \
+		--input-dirs "$(find_dirs_containing_comment_tags "+k8s:conversion-gen=")" \
 		--output-file-base zz_generated.conversion

hack/update-generated-deep-copies.sh

@@ -1,10 +1,11 @@
 #!/bin/bash
 source "$(dirname "${BASH_SOURCE}")/lib/init.sh"
+source "$(dirname "${BASH_SOURCE}")/lib/generator-help.sh"
 
 go build -o "${OS_OUTPUT_BINPATH}/deepcopy-gen" "k8s.io/code-generator/cmd/deepcopy-gen"
 
 ${OS_OUTPUT_BINPATH}/deepcopy-gen \
                 --go-header-file "hack/boilerplate/boilerplate.go.txt" \
-                --input-dirs "${PRJ_PREFIX}/pkg/apis/componentconfig,${PRJ_PREFIX}/pkg/apis/componentconfig/v1alpha1,${PRJ_PREFIX}/pkg/api,${PRJ_PREFIX}/pkg/api/v1alpha1" \
+                --input-dirs "$(find_dirs_containing_comment_tags "+k8s:deepcopy-gen=")" \
                 --output-file-base zz_generated.deepcopy
 

hack/update-generated-defaulters.sh

@@ -1,10 +1,11 @@
 #!/bin/bash
 source "$(dirname "${BASH_SOURCE}")/lib/init.sh"
+source "$(dirname "${BASH_SOURCE}")/lib/generator-help.sh"
 
 go build -o "${OS_OUTPUT_BINPATH}/defaulter-gen" "k8s.io/code-generator/cmd/defaulter-gen"
 
 ${OS_OUTPUT_BINPATH}/defaulter-gen \
                 --go-header-file "hack/boilerplate/boilerplate.go.txt" \
-                --input-dirs "${PRJ_PREFIX}/pkg/apis/componentconfig/v1alpha1,${PRJ_PREFIX}/pkg/api/v1alpha1" \
+                --input-dirs "$(find_dirs_containing_comment_tags "+k8s:defaulter-gen=")" \
 		--extra-peer-dirs "${PRJ_PREFIX}/pkg/apis/componentconfig/v1alpha1,${PRJ_PREFIX}/pkg/api/v1alpha1" \
                 --output-file-base zz_generated.defaults

hack/update-gofmt.sh

@@ -23,7 +23,7 @@ DESCHEDULER_ROOT=$(dirname "${BASH_SOURCE}")/..
 
 GO_VERSION=($(go version))
 
-if [[ -z $(echo "${GO_VERSION[2]}" | grep -E 'go1.14|go1.15|go1.16') ]]; then
+if [[ -z $(echo "${GO_VERSION[2]}" | grep -E 'go1.17|go1.18|go1.19') ]]; then
   echo "Unknown go version '${GO_VERSION[2]}', skipping gofmt."
   exit 1
 fi

hack/verify-chart.sh

@@ -0,0 +1 @@
+${CONTAINER_ENGINE:-docker} run -it --rm --network host --workdir=/data --volume ~/.kube/config:/root/.kube/config:ro --volume $(pwd):/data quay.io/helmpack/chart-testing:v3.7.0 /bin/bash -c "git config --global --add safe.directory /data; ct install --config=.github/ci/ct.yaml --helm-extra-set-args=\"--set=kind=Deployment\""

hack/verify-conversions.sh

@@ -16,12 +16,7 @@ git archive --format=tar --prefix=descheduler/ "$(git write-tree)" | (cd "${_des
 _deschedulertmp="${_deschedulertmp}/descheduler"
 
 pushd "${_deschedulertmp}" > /dev/null 2>&1
-go build -o "${OS_OUTPUT_BINPATH}/conversion-gen" "k8s.io/code-generator/cmd/conversion-gen"
-
-${OS_OUTPUT_BINPATH}/conversion-gen \
-		--go-header-file "hack/boilerplate/boilerplate.go.txt" \
-		--input-dirs "./pkg/apis/componentconfig/v1alpha1,./pkg/api/v1alpha1" \
-		--output-file-base zz_generated.conversion
+./hack/update-generated-conversions.sh
 popd > /dev/null 2>&1
 
 pushd "${DESCHEDULER_ROOT}" > /dev/null 2>&1

hack/verify-deep-copies.sh

@@ -16,12 +16,7 @@ git archive --format=tar --prefix=descheduler/ "$(git write-tree)" | (cd "${_des
 _deschedulertmp="${_deschedulertmp}/descheduler"
 
 pushd "${_deschedulertmp}" > /dev/null 2>&1
-go build -o "${OS_OUTPUT_BINPATH}/deepcopy-gen" "k8s.io/code-generator/cmd/deepcopy-gen"
-
-${OS_OUTPUT_BINPATH}/deepcopy-gen \
-                --go-header-file "hack/boilerplate/boilerplate.go.txt" \
-                --input-dirs "./pkg/apis/componentconfig,./pkg/apis/componentconfig/v1alpha1,./pkg/api,./pkg/api/v1alpha1" \
-                --output-file-base zz_generated.deepcopy
+./hack/update-generated-deep-copies.sh
 popd > /dev/null 2>&1
 
 pushd "${DESCHEDULER_ROOT}" > /dev/null 2>&1

hack/verify-defaulters.sh

@@ -15,13 +15,7 @@ git archive --format=tar --prefix=descheduler/ "$(git write-tree)" | (cd "${_des
 _deschedulertmp="${_deschedulertmp}/descheduler"
 
 pushd "${_deschedulertmp}" > /dev/null 2>&1
-go build -o "${OS_OUTPUT_BINPATH}/defaulter-gen" "k8s.io/code-generator/cmd/defaulter-gen"
-
-${OS_OUTPUT_BINPATH}/defaulter-gen \
-            --go-header-file "hack/boilerplate/boilerplate.go.txt" \
-            --input-dirs "${PRJ_PREFIX}/pkg/apis/componentconfig/v1alpha1,${PRJ_PREFIX}/pkg/api/v1alpha1" \
-            --extra-peer-dirs "${PRJ_PREFIX}/pkg/apis/componentconfig/v1alpha1,${PRJ_PREFIX}/pkg/api/v1alpha1" \
-            --output-file-base zz_generated.defaults
+./hack/update-generated-defaulters.sh
 popd > /dev/null 2>&1
 
 pushd "${DESCHEDULER_ROOT}" > /dev/null 2>&1

hack/verify-gofmt.sh

@@ -23,7 +23,7 @@ DESCHEDULER_ROOT=$(dirname "${BASH_SOURCE}")/..
 
 GO_VERSION=($(go version))
 
-if [[ -z $(echo "${GO_VERSION[2]}" | grep -E 'go1.14|go1.15|go1.16') ]]; then
+if [[ -z $(echo "${GO_VERSION[2]}" | grep -E 'go1.17|go1.18|go1.19') ]]; then
   echo "Unknown go version '${GO_VERSION[2]}', skipping gofmt."
   exit 1
 fi

kubernetes/base/rbac.yaml

@@ -4,7 +4,7 @@ apiVersion: rbac.authorization.k8s.io/v1
 metadata:
   name: descheduler-cluster-role
 rules:
-- apiGroups: [""]
+- apiGroups: ["events.k8s.io"]
   resources: ["events"]
   verbs: ["create", "update"]
 - apiGroups: [""]
@@ -12,7 +12,7 @@ rules:
   verbs: ["get", "watch", "list"]
 - apiGroups: [""]
   resources: ["namespaces"]
-  verbs: ["get", "list"]
+  verbs: ["get", "watch", "list"]
 - apiGroups: [""]
   resources: ["pods"]
   verbs: ["get", "watch", "list", "delete"]
@@ -22,6 +22,13 @@ rules:
 - apiGroups: ["scheduling.k8s.io"]
   resources: ["priorityclasses"]
   verbs: ["get", "watch", "list"]
+- apiGroups: ["coordination.k8s.io"]
+  resources: ["leases"]
+  verbs: ["create"]
+- apiGroups: ["coordination.k8s.io"]
+  resources: ["leases"]
+  resourceNames: ["descheduler"]
+  verbs: ["get", "patch", "delete"]
 ---
 apiVersion: v1
 kind: ServiceAccount
@@ -41,4 +48,3 @@ subjects:
   - name: descheduler-sa
     kind: ServiceAccount
     namespace: kube-system
-

kubernetes/cronjob/cronjob.yaml

@@ -1,5 +1,5 @@
 ---
-apiVersion: batch/v1  # for k8s version < 1.21.0, use batch/v1beta1
+apiVersion: batch/v1
 kind: CronJob
 metadata:
   name: descheduler-cronjob
@@ -16,7 +16,7 @@ spec:
           priorityClassName: system-cluster-critical
           containers:
           - name: descheduler
-            image: k8s.gcr.io/descheduler/descheduler:v0.22.0
+            image: registry.k8s.io/descheduler/descheduler:v0.25.1
             volumeMounts:
             - mountPath: /policy-dir
               name: policy-volume
@@ -31,6 +31,14 @@ spec:
               requests:
                 cpu: "500m"
                 memory: "256Mi"
+            livenessProbe:
+              failureThreshold: 3
+              httpGet:
+                path: /healthz
+                port: 10258
+                scheme: HTTPS
+              initialDelaySeconds: 3
+              periodSeconds: 10
             securityContext:
               allowPrivilegeEscalation: false
               capabilities:

kubernetes/deployment/deployment.yaml

@@ -19,7 +19,7 @@ spec:
       serviceAccountName: descheduler-sa
       containers:
         - name: descheduler
-          image: k8s.gcr.io/descheduler/descheduler:v0.22.0
+          image: registry.k8s.io/descheduler/descheduler:v0.25.1
           imagePullPolicy: IfNotPresent
           command:
             - "/bin/descheduler"
@@ -33,6 +33,14 @@ spec:
           ports:
           - containerPort: 10258
             protocol: TCP
+          livenessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /healthz
+              port: 10258
+              scheme: HTTPS
+            initialDelaySeconds: 3
+            periodSeconds: 10
           resources:
             requests:
               cpu: 500m

kubernetes/job/job.yaml

@@ -14,7 +14,7 @@ spec:
       priorityClassName: system-cluster-critical
       containers:
         - name: descheduler
-          image: k8s.gcr.io/descheduler/descheduler:v0.22.0
+          image: registry.k8s.io/descheduler/descheduler:v0.25.1
           volumeMounts:
           - mountPath: /policy-dir
             name: policy-volume
@@ -29,6 +29,14 @@ spec:
             requests:
               cpu: "500m"
               memory: "256Mi"
+          livenessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /healthz
+              port: 10258
+              scheme: HTTPS
+            initialDelaySeconds: 3
+            periodSeconds: 10
           securityContext:
             allowPrivilegeEscalation: false
             capabilities:

metrics/metrics.go

@@ -34,16 +34,16 @@ var (
 		&metrics.CounterOpts{
 			Subsystem:      DeschedulerSubsystem,
 			Name:           "pods_evicted",
-			Help:           "Number of evicted pods, by the result, by the strategy, by the namespace. 'failed' result means a pod could not be evicted",
+			Help:           "Number of evicted pods, by the result, by the strategy, by the namespace, by the node name. 'error' result means a pod could not be evicted",
 			StabilityLevel: metrics.ALPHA,
-		}, []string{"result", "strategy", "namespace"})
+		}, []string{"result", "strategy", "namespace", "node"})
 
 	buildInfo = metrics.NewGauge(
 		&metrics.GaugeOpts{
 			Subsystem:      DeschedulerSubsystem,
 			Name:           "build_info",
 			Help:           "Build info about descheduler, including Go version, Descheduler version, Git SHA, Git branch",
-			ConstLabels:    map[string]string{"GoVersion": version.Get().GoVersion, "DeschedulerVersion": version.Get().GitVersion, "GitBranch": version.Get().GitBranch, "GitSha1": version.Get().GitSha1},
+			ConstLabels:    map[string]string{"GoVersion": version.Get().GoVersion, "AppVersion": version.Get().Major + "." + version.Get().Minor, "DeschedulerVersion": version.Get().GitVersion, "GitBranch": version.Get().GitBranch, "GitSha1": version.Get().GitSha1},
 			StabilityLevel: metrics.ALPHA,
 		},
 	)

pkg/api/sort.go

@@ -0,0 +1,23 @@
+/*
+Copyright 2017 The Kubernetes Authors.
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+    http://www.apache.org/licenses/LICENSE-2.0
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package api
+
+import "sort"
+
+func SortProfilesByName(profiles []Profile) []Profile {
+	sort.Slice(profiles, func(i, j int) bool {
+		return profiles[i].Name < profiles[j].Name
+	})
+	return profiles
+}

pkg/api/types.go

@@ -19,6 +19,7 @@ package api
 import (
 	v1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/runtime"
 )
 
 // +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
@@ -26,37 +27,17 @@ import (
 type DeschedulerPolicy struct {
 	metav1.TypeMeta
 
-	// Strategies
-	Strategies StrategyList
+	// Profiles
+	Profiles []Profile
 
 	// NodeSelector for a set of nodes to operate over
 	NodeSelector *string
 
-	// EvictLocalStoragePods allows pods using local storage to be evicted.
-	EvictLocalStoragePods *bool
-
-	// EvictSystemCriticalPods allows eviction of pods of any priority (including Kubernetes system pods)
-	EvictSystemCriticalPods *bool
-
-	// IgnorePVCPods prevents pods with PVCs from being evicted.
-	IgnorePVCPods *bool
-
 	// MaxNoOfPodsToEvictPerNode restricts maximum of pods to be evicted per node.
-	MaxNoOfPodsToEvictPerNode *int
-}
+	MaxNoOfPodsToEvictPerNode *uint
 
-type StrategyName string
-type StrategyList map[StrategyName]DeschedulerStrategy
-
-type DeschedulerStrategy struct {
-	// Enabled or disabled
-	Enabled bool
-
-	// Weight
-	Weight int
-
-	// Strategy parameters
-	Params *StrategyParameters
+	// MaxNoOfPodsToEvictPerNamespace restricts maximum of pods to be evicted per namespace.
+	MaxNoOfPodsToEvictPerNamespace *uint
 }
 
 // Namespaces carries a list of included/excluded namespaces
@@ -66,50 +47,38 @@ type Namespaces struct {
 	Exclude []string
 }
 
-// Besides Namespaces only one of its members may be specified
-// TODO(jchaloup): move Namespaces ThresholdPriority and ThresholdPriorityClassName to individual strategies
-//  once the policy version is bumped to v1alpha2
-type StrategyParameters struct {
-	NodeResourceUtilizationThresholds *NodeResourceUtilizationThresholds
-	NodeAffinityType                  []string
-	PodsHavingTooManyRestarts         *PodsHavingTooManyRestarts
-	PodLifeTime                       *PodLifeTime
-	RemoveDuplicates                  *RemoveDuplicates
-	FailedPods                        *FailedPods
-	IncludeSoftConstraints            bool
-	Namespaces                        *Namespaces
-	ThresholdPriority                 *int32
-	ThresholdPriorityClassName        string
-	LabelSelector                     *metav1.LabelSelector
-	NodeFit                           bool
+type (
+	Percentage         float64
+	ResourceThresholds map[v1.ResourceName]Percentage
+)
+
+type PriorityThreshold struct {
+	Value *int32
+	Name  string
 }
 
-type Percentage float64
-type ResourceThresholds map[v1.ResourceName]Percentage
-
-type NodeResourceUtilizationThresholds struct {
-	Thresholds       ResourceThresholds
-	TargetThresholds ResourceThresholds
-	NumberOfNodes    int
+type Profile struct {
+	Name          string
+	PluginConfigs []PluginConfig
+	Plugins       Plugins
 }
 
-type PodsHavingTooManyRestarts struct {
-	PodRestartThreshold     int32
-	IncludingInitContainers bool
+type PluginConfig struct {
+	Name string
+	Args runtime.Object
 }
 
-type RemoveDuplicates struct {
-	ExcludeOwnerKinds []string
+type Plugins struct {
+	PreSort           PluginSet
+	Sort              PluginSet
+	Deschedule        PluginSet
+	Balance           PluginSet
+	Evict             PluginSet
+	Filter            PluginSet
+	PreEvictionFilter PluginSet
 }
 
-type PodLifeTime struct {
-	MaxPodLifeTimeSeconds *uint
-	PodStatusPhases       []string
-}
-
-type FailedPods struct {
-	ExcludeOwnerKinds       []string
-	MinPodLifetimeSeconds   *uint
-	Reasons                 []string
-	IncludingInitContainers bool
+type PluginSet struct {
+	Enabled  []string
+	Disabled []string
 }

pkg/api/v1alpha1/doc.go

@@ -15,7 +15,6 @@ limitations under the License.
 */
 
 // +k8s:deepcopy-gen=package,register
-// +k8s:conversion-gen=sigs.k8s.io/descheduler/pkg/api
 // +k8s:defaulter-gen=TypeMeta
 
 // Package v1alpha1 is the v1alpha1 version of the descheduler API

pkg/api/v1alpha1/register.go

@@ -28,8 +28,10 @@ var (
 )
 
 // GroupName is the group name used in this package
-const GroupName = "descheduler"
-const GroupVersion = "v1alpha1"
+const (
+	GroupName    = "descheduler"
+	GroupVersion = "v1alpha1"
+)
 
 // SchemeGroupVersion is group version used to register these objects
 var SchemeGroupVersion = schema.GroupVersion{Group: GroupName, Version: GroupVersion}

pkg/api/v1alpha1/types.go

@@ -32,6 +32,9 @@ type DeschedulerPolicy struct {
 	// NodeSelector for a set of nodes to operate over
 	NodeSelector *string `json:"nodeSelector,omitempty"`
 
+	// EvictFailedBarePods allows pods without ownerReferences and in failed phase to be evicted.
+	EvictFailedBarePods *bool `json:"evictFailedBarePods,omitempty"`
+
 	// EvictLocalStoragePods allows pods using local storage to be evicted.
 	EvictLocalStoragePods *bool `json:"evictLocalStoragePods,omitempty"`
 
@@ -42,11 +45,16 @@ type DeschedulerPolicy struct {
 	IgnorePVCPods *bool `json:"ignorePvcPods,omitempty"`
 
 	// MaxNoOfPodsToEvictPerNode restricts maximum of pods to be evicted per node.
-	MaxNoOfPodsToEvictPerNode *int `json:"maxNoOfPodsToEvictPerNode,omitempty"`
+	MaxNoOfPodsToEvictPerNode *uint `json:"maxNoOfPodsToEvictPerNode,omitempty"`
+
+	// MaxNoOfPodsToEvictPerNamespace restricts maximum of pods to be evicted per namespace.
+	MaxNoOfPodsToEvictPerNamespace *uint `json:"maxNoOfPodsToEvictPerNamespace,omitempty"`
 }
 
-type StrategyName string
-type StrategyList map[StrategyName]DeschedulerStrategy
+type (
+	StrategyName string
+	StrategyList map[StrategyName]DeschedulerStrategy
+)
 
 type DeschedulerStrategy struct {
 	// Enabled or disabled
@@ -80,15 +88,20 @@ type StrategyParameters struct {
 	ThresholdPriorityClassName        string                             `json:"thresholdPriorityClassName"`
 	LabelSelector                     *metav1.LabelSelector              `json:"labelSelector"`
 	NodeFit                           bool                               `json:"nodeFit"`
+	IncludePreferNoSchedule           bool                               `json:"includePreferNoSchedule"`
+	ExcludedTaints                    []string                           `json:"excludedTaints,omitempty"`
 }
 
-type Percentage float64
-type ResourceThresholds map[v1.ResourceName]Percentage
+type (
+	Percentage         float64
+	ResourceThresholds map[v1.ResourceName]Percentage
+)
 
 type NodeResourceUtilizationThresholds struct {
-	Thresholds       ResourceThresholds `json:"thresholds,omitempty"`
-	TargetThresholds ResourceThresholds `json:"targetThresholds,omitempty"`
-	NumberOfNodes    int                `json:"numberOfNodes,omitempty"`
+	UseDeviationThresholds bool               `json:"useDeviationThresholds,omitempty"`
+	Thresholds             ResourceThresholds `json:"thresholds,omitempty"`
+	TargetThresholds       ResourceThresholds `json:"targetThresholds,omitempty"`
+	NumberOfNodes          int                `json:"numberOfNodes,omitempty"`
 }
 
 type PodsHavingTooManyRestarts struct {
@@ -102,7 +115,10 @@ type RemoveDuplicates struct {
 
 type PodLifeTime struct {
 	MaxPodLifeTimeSeconds *uint    `json:"maxPodLifeTimeSeconds,omitempty"`
-	PodStatusPhases       []string `json:"podStatusPhases,omitempty"`
+	States                []string `json:"states,omitempty"`
+
+	// Deprecated: Use States instead.
+	PodStatusPhases []string `json:"podStatusPhases,omitempty"`
 }
 
 type FailedPods struct {

pkg/api/v1alpha1/zz_generated.conversion.go

@@ -1,362 +0,0 @@
-// +build !ignore_autogenerated
-
-/*
-Copyright 2021 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-// Code generated by conversion-gen. DO NOT EDIT.
-
-package v1alpha1
-
-import (
-	unsafe "unsafe"
-
-	v1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	conversion "k8s.io/apimachinery/pkg/conversion"
-	runtime "k8s.io/apimachinery/pkg/runtime"
-	api "sigs.k8s.io/descheduler/pkg/api"
-)
-
-func init() {
-	localSchemeBuilder.Register(RegisterConversions)
-}
-
-// RegisterConversions adds conversion functions to the given scheme.
-// Public to allow building arbitrary schemes.
-func RegisterConversions(s *runtime.Scheme) error {
-	if err := s.AddGeneratedConversionFunc((*DeschedulerPolicy)(nil), (*api.DeschedulerPolicy)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_DeschedulerPolicy_To_api_DeschedulerPolicy(a.(*DeschedulerPolicy), b.(*api.DeschedulerPolicy), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.DeschedulerPolicy)(nil), (*DeschedulerPolicy)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_DeschedulerPolicy_To_v1alpha1_DeschedulerPolicy(a.(*api.DeschedulerPolicy), b.(*DeschedulerPolicy), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*DeschedulerStrategy)(nil), (*api.DeschedulerStrategy)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_DeschedulerStrategy_To_api_DeschedulerStrategy(a.(*DeschedulerStrategy), b.(*api.DeschedulerStrategy), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.DeschedulerStrategy)(nil), (*DeschedulerStrategy)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_DeschedulerStrategy_To_v1alpha1_DeschedulerStrategy(a.(*api.DeschedulerStrategy), b.(*DeschedulerStrategy), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*FailedPods)(nil), (*api.FailedPods)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_FailedPods_To_api_FailedPods(a.(*FailedPods), b.(*api.FailedPods), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.FailedPods)(nil), (*FailedPods)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_FailedPods_To_v1alpha1_FailedPods(a.(*api.FailedPods), b.(*FailedPods), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*Namespaces)(nil), (*api.Namespaces)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_Namespaces_To_api_Namespaces(a.(*Namespaces), b.(*api.Namespaces), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.Namespaces)(nil), (*Namespaces)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_Namespaces_To_v1alpha1_Namespaces(a.(*api.Namespaces), b.(*Namespaces), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*NodeResourceUtilizationThresholds)(nil), (*api.NodeResourceUtilizationThresholds)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_NodeResourceUtilizationThresholds_To_api_NodeResourceUtilizationThresholds(a.(*NodeResourceUtilizationThresholds), b.(*api.NodeResourceUtilizationThresholds), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.NodeResourceUtilizationThresholds)(nil), (*NodeResourceUtilizationThresholds)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds(a.(*api.NodeResourceUtilizationThresholds), b.(*NodeResourceUtilizationThresholds), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*PodLifeTime)(nil), (*api.PodLifeTime)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_PodLifeTime_To_api_PodLifeTime(a.(*PodLifeTime), b.(*api.PodLifeTime), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.PodLifeTime)(nil), (*PodLifeTime)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_PodLifeTime_To_v1alpha1_PodLifeTime(a.(*api.PodLifeTime), b.(*PodLifeTime), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*PodsHavingTooManyRestarts)(nil), (*api.PodsHavingTooManyRestarts)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_PodsHavingTooManyRestarts_To_api_PodsHavingTooManyRestarts(a.(*PodsHavingTooManyRestarts), b.(*api.PodsHavingTooManyRestarts), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.PodsHavingTooManyRestarts)(nil), (*PodsHavingTooManyRestarts)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_PodsHavingTooManyRestarts_To_v1alpha1_PodsHavingTooManyRestarts(a.(*api.PodsHavingTooManyRestarts), b.(*PodsHavingTooManyRestarts), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*RemoveDuplicates)(nil), (*api.RemoveDuplicates)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_RemoveDuplicates_To_api_RemoveDuplicates(a.(*RemoveDuplicates), b.(*api.RemoveDuplicates), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.RemoveDuplicates)(nil), (*RemoveDuplicates)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_RemoveDuplicates_To_v1alpha1_RemoveDuplicates(a.(*api.RemoveDuplicates), b.(*RemoveDuplicates), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*StrategyParameters)(nil), (*api.StrategyParameters)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_v1alpha1_StrategyParameters_To_api_StrategyParameters(a.(*StrategyParameters), b.(*api.StrategyParameters), scope)
-	}); err != nil {
-		return err
-	}
-	if err := s.AddGeneratedConversionFunc((*api.StrategyParameters)(nil), (*StrategyParameters)(nil), func(a, b interface{}, scope conversion.Scope) error {
-		return Convert_api_StrategyParameters_To_v1alpha1_StrategyParameters(a.(*api.StrategyParameters), b.(*StrategyParameters), scope)
-	}); err != nil {
-		return err
-	}
-	return nil
-}
-
-func autoConvert_v1alpha1_DeschedulerPolicy_To_api_DeschedulerPolicy(in *DeschedulerPolicy, out *api.DeschedulerPolicy, s conversion.Scope) error {
-	out.Strategies = *(*api.StrategyList)(unsafe.Pointer(&in.Strategies))
-	out.NodeSelector = (*string)(unsafe.Pointer(in.NodeSelector))
-	out.EvictLocalStoragePods = (*bool)(unsafe.Pointer(in.EvictLocalStoragePods))
-	out.EvictSystemCriticalPods = (*bool)(unsafe.Pointer(in.EvictSystemCriticalPods))
-	out.IgnorePVCPods = (*bool)(unsafe.Pointer(in.IgnorePVCPods))
-	out.MaxNoOfPodsToEvictPerNode = (*int)(unsafe.Pointer(in.MaxNoOfPodsToEvictPerNode))
-	return nil
-}
-
-// Convert_v1alpha1_DeschedulerPolicy_To_api_DeschedulerPolicy is an autogenerated conversion function.
-func Convert_v1alpha1_DeschedulerPolicy_To_api_DeschedulerPolicy(in *DeschedulerPolicy, out *api.DeschedulerPolicy, s conversion.Scope) error {
-	return autoConvert_v1alpha1_DeschedulerPolicy_To_api_DeschedulerPolicy(in, out, s)
-}
-
-func autoConvert_api_DeschedulerPolicy_To_v1alpha1_DeschedulerPolicy(in *api.DeschedulerPolicy, out *DeschedulerPolicy, s conversion.Scope) error {
-	out.Strategies = *(*StrategyList)(unsafe.Pointer(&in.Strategies))
-	out.NodeSelector = (*string)(unsafe.Pointer(in.NodeSelector))
-	out.EvictLocalStoragePods = (*bool)(unsafe.Pointer(in.EvictLocalStoragePods))
-	out.EvictSystemCriticalPods = (*bool)(unsafe.Pointer(in.EvictSystemCriticalPods))
-	out.IgnorePVCPods = (*bool)(unsafe.Pointer(in.IgnorePVCPods))
-	out.MaxNoOfPodsToEvictPerNode = (*int)(unsafe.Pointer(in.MaxNoOfPodsToEvictPerNode))
-	return nil
-}
-
-// Convert_api_DeschedulerPolicy_To_v1alpha1_DeschedulerPolicy is an autogenerated conversion function.
-func Convert_api_DeschedulerPolicy_To_v1alpha1_DeschedulerPolicy(in *api.DeschedulerPolicy, out *DeschedulerPolicy, s conversion.Scope) error {
-	return autoConvert_api_DeschedulerPolicy_To_v1alpha1_DeschedulerPolicy(in, out, s)
-}
-
-func autoConvert_v1alpha1_DeschedulerStrategy_To_api_DeschedulerStrategy(in *DeschedulerStrategy, out *api.DeschedulerStrategy, s conversion.Scope) error {
-	out.Enabled = in.Enabled
-	out.Weight = in.Weight
-	out.Params = (*api.StrategyParameters)(unsafe.Pointer(in.Params))
-	return nil
-}
-
-// Convert_v1alpha1_DeschedulerStrategy_To_api_DeschedulerStrategy is an autogenerated conversion function.
-func Convert_v1alpha1_DeschedulerStrategy_To_api_DeschedulerStrategy(in *DeschedulerStrategy, out *api.DeschedulerStrategy, s conversion.Scope) error {
-	return autoConvert_v1alpha1_DeschedulerStrategy_To_api_DeschedulerStrategy(in, out, s)
-}
-
-func autoConvert_api_DeschedulerStrategy_To_v1alpha1_DeschedulerStrategy(in *api.DeschedulerStrategy, out *DeschedulerStrategy, s conversion.Scope) error {
-	out.Enabled = in.Enabled
-	out.Weight = in.Weight
-	out.Params = (*StrategyParameters)(unsafe.Pointer(in.Params))
-	return nil
-}
-
-// Convert_api_DeschedulerStrategy_To_v1alpha1_DeschedulerStrategy is an autogenerated conversion function.
-func Convert_api_DeschedulerStrategy_To_v1alpha1_DeschedulerStrategy(in *api.DeschedulerStrategy, out *DeschedulerStrategy, s conversion.Scope) error {
-	return autoConvert_api_DeschedulerStrategy_To_v1alpha1_DeschedulerStrategy(in, out, s)
-}
-
-func autoConvert_v1alpha1_FailedPods_To_api_FailedPods(in *FailedPods, out *api.FailedPods, s conversion.Scope) error {
-	out.ExcludeOwnerKinds = *(*[]string)(unsafe.Pointer(&in.ExcludeOwnerKinds))
-	out.MinPodLifetimeSeconds = (*uint)(unsafe.Pointer(in.MinPodLifetimeSeconds))
-	out.Reasons = *(*[]string)(unsafe.Pointer(&in.Reasons))
-	out.IncludingInitContainers = in.IncludingInitContainers
-	return nil
-}
-
-// Convert_v1alpha1_FailedPods_To_api_FailedPods is an autogenerated conversion function.
-func Convert_v1alpha1_FailedPods_To_api_FailedPods(in *FailedPods, out *api.FailedPods, s conversion.Scope) error {
-	return autoConvert_v1alpha1_FailedPods_To_api_FailedPods(in, out, s)
-}
-
-func autoConvert_api_FailedPods_To_v1alpha1_FailedPods(in *api.FailedPods, out *FailedPods, s conversion.Scope) error {
-	out.ExcludeOwnerKinds = *(*[]string)(unsafe.Pointer(&in.ExcludeOwnerKinds))
-	out.MinPodLifetimeSeconds = (*uint)(unsafe.Pointer(in.MinPodLifetimeSeconds))
-	out.Reasons = *(*[]string)(unsafe.Pointer(&in.Reasons))
-	out.IncludingInitContainers = in.IncludingInitContainers
-	return nil
-}
-
-// Convert_api_FailedPods_To_v1alpha1_FailedPods is an autogenerated conversion function.
-func Convert_api_FailedPods_To_v1alpha1_FailedPods(in *api.FailedPods, out *FailedPods, s conversion.Scope) error {
-	return autoConvert_api_FailedPods_To_v1alpha1_FailedPods(in, out, s)
-}
-
-func autoConvert_v1alpha1_Namespaces_To_api_Namespaces(in *Namespaces, out *api.Namespaces, s conversion.Scope) error {
-	out.Include = *(*[]string)(unsafe.Pointer(&in.Include))
-	out.Exclude = *(*[]string)(unsafe.Pointer(&in.Exclude))
-	return nil
-}
-
-// Convert_v1alpha1_Namespaces_To_api_Namespaces is an autogenerated conversion function.
-func Convert_v1alpha1_Namespaces_To_api_Namespaces(in *Namespaces, out *api.Namespaces, s conversion.Scope) error {
-	return autoConvert_v1alpha1_Namespaces_To_api_Namespaces(in, out, s)
-}
-
-func autoConvert_api_Namespaces_To_v1alpha1_Namespaces(in *api.Namespaces, out *Namespaces, s conversion.Scope) error {
-	out.Include = *(*[]string)(unsafe.Pointer(&in.Include))
-	out.Exclude = *(*[]string)(unsafe.Pointer(&in.Exclude))
-	return nil
-}
-
-// Convert_api_Namespaces_To_v1alpha1_Namespaces is an autogenerated conversion function.
-func Convert_api_Namespaces_To_v1alpha1_Namespaces(in *api.Namespaces, out *Namespaces, s conversion.Scope) error {
-	return autoConvert_api_Namespaces_To_v1alpha1_Namespaces(in, out, s)
-}
-
-func autoConvert_v1alpha1_NodeResourceUtilizationThresholds_To_api_NodeResourceUtilizationThresholds(in *NodeResourceUtilizationThresholds, out *api.NodeResourceUtilizationThresholds, s conversion.Scope) error {
-	out.Thresholds = *(*api.ResourceThresholds)(unsafe.Pointer(&in.Thresholds))
-	out.TargetThresholds = *(*api.ResourceThresholds)(unsafe.Pointer(&in.TargetThresholds))
-	out.NumberOfNodes = in.NumberOfNodes
-	return nil
-}
-
-// Convert_v1alpha1_NodeResourceUtilizationThresholds_To_api_NodeResourceUtilizationThresholds is an autogenerated conversion function.
-func Convert_v1alpha1_NodeResourceUtilizationThresholds_To_api_NodeResourceUtilizationThresholds(in *NodeResourceUtilizationThresholds, out *api.NodeResourceUtilizationThresholds, s conversion.Scope) error {
-	return autoConvert_v1alpha1_NodeResourceUtilizationThresholds_To_api_NodeResourceUtilizationThresholds(in, out, s)
-}
-
-func autoConvert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds(in *api.NodeResourceUtilizationThresholds, out *NodeResourceUtilizationThresholds, s conversion.Scope) error {
-	out.Thresholds = *(*ResourceThresholds)(unsafe.Pointer(&in.Thresholds))
-	out.TargetThresholds = *(*ResourceThresholds)(unsafe.Pointer(&in.TargetThresholds))
-	out.NumberOfNodes = in.NumberOfNodes
-	return nil
-}
-
-// Convert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds is an autogenerated conversion function.
-func Convert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds(in *api.NodeResourceUtilizationThresholds, out *NodeResourceUtilizationThresholds, s conversion.Scope) error {
-	return autoConvert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds(in, out, s)
-}
-
-func autoConvert_v1alpha1_PodLifeTime_To_api_PodLifeTime(in *PodLifeTime, out *api.PodLifeTime, s conversion.Scope) error {
-	out.MaxPodLifeTimeSeconds = (*uint)(unsafe.Pointer(in.MaxPodLifeTimeSeconds))
-	out.PodStatusPhases = *(*[]string)(unsafe.Pointer(&in.PodStatusPhases))
-	return nil
-}
-
-// Convert_v1alpha1_PodLifeTime_To_api_PodLifeTime is an autogenerated conversion function.
-func Convert_v1alpha1_PodLifeTime_To_api_PodLifeTime(in *PodLifeTime, out *api.PodLifeTime, s conversion.Scope) error {
-	return autoConvert_v1alpha1_PodLifeTime_To_api_PodLifeTime(in, out, s)
-}
-
-func autoConvert_api_PodLifeTime_To_v1alpha1_PodLifeTime(in *api.PodLifeTime, out *PodLifeTime, s conversion.Scope) error {
-	out.MaxPodLifeTimeSeconds = (*uint)(unsafe.Pointer(in.MaxPodLifeTimeSeconds))
-	out.PodStatusPhases = *(*[]string)(unsafe.Pointer(&in.PodStatusPhases))
-	return nil
-}
-
-// Convert_api_PodLifeTime_To_v1alpha1_PodLifeTime is an autogenerated conversion function.
-func Convert_api_PodLifeTime_To_v1alpha1_PodLifeTime(in *api.PodLifeTime, out *PodLifeTime, s conversion.Scope) error {
-	return autoConvert_api_PodLifeTime_To_v1alpha1_PodLifeTime(in, out, s)
-}
-
-func autoConvert_v1alpha1_PodsHavingTooManyRestarts_To_api_PodsHavingTooManyRestarts(in *PodsHavingTooManyRestarts, out *api.PodsHavingTooManyRestarts, s conversion.Scope) error {
-	out.PodRestartThreshold = in.PodRestartThreshold
-	out.IncludingInitContainers = in.IncludingInitContainers
-	return nil
-}
-
-// Convert_v1alpha1_PodsHavingTooManyRestarts_To_api_PodsHavingTooManyRestarts is an autogenerated conversion function.
-func Convert_v1alpha1_PodsHavingTooManyRestarts_To_api_PodsHavingTooManyRestarts(in *PodsHavingTooManyRestarts, out *api.PodsHavingTooManyRestarts, s conversion.Scope) error {
-	return autoConvert_v1alpha1_PodsHavingTooManyRestarts_To_api_PodsHavingTooManyRestarts(in, out, s)
-}
-
-func autoConvert_api_PodsHavingTooManyRestarts_To_v1alpha1_PodsHavingTooManyRestarts(in *api.PodsHavingTooManyRestarts, out *PodsHavingTooManyRestarts, s conversion.Scope) error {
-	out.PodRestartThreshold = in.PodRestartThreshold
-	out.IncludingInitContainers = in.IncludingInitContainers
-	return nil
-}
-
-// Convert_api_PodsHavingTooManyRestarts_To_v1alpha1_PodsHavingTooManyRestarts is an autogenerated conversion function.
-func Convert_api_PodsHavingTooManyRestarts_To_v1alpha1_PodsHavingTooManyRestarts(in *api.PodsHavingTooManyRestarts, out *PodsHavingTooManyRestarts, s conversion.Scope) error {
-	return autoConvert_api_PodsHavingTooManyRestarts_To_v1alpha1_PodsHavingTooManyRestarts(in, out, s)
-}
-
-func autoConvert_v1alpha1_RemoveDuplicates_To_api_RemoveDuplicates(in *RemoveDuplicates, out *api.RemoveDuplicates, s conversion.Scope) error {
-	out.ExcludeOwnerKinds = *(*[]string)(unsafe.Pointer(&in.ExcludeOwnerKinds))
-	return nil
-}
-
-// Convert_v1alpha1_RemoveDuplicates_To_api_RemoveDuplicates is an autogenerated conversion function.
-func Convert_v1alpha1_RemoveDuplicates_To_api_RemoveDuplicates(in *RemoveDuplicates, out *api.RemoveDuplicates, s conversion.Scope) error {
-	return autoConvert_v1alpha1_RemoveDuplicates_To_api_RemoveDuplicates(in, out, s)
-}
-
-func autoConvert_api_RemoveDuplicates_To_v1alpha1_RemoveDuplicates(in *api.RemoveDuplicates, out *RemoveDuplicates, s conversion.Scope) error {
-	out.ExcludeOwnerKinds = *(*[]string)(unsafe.Pointer(&in.ExcludeOwnerKinds))
-	return nil
-}
-
-// Convert_api_RemoveDuplicates_To_v1alpha1_RemoveDuplicates is an autogenerated conversion function.
-func Convert_api_RemoveDuplicates_To_v1alpha1_RemoveDuplicates(in *api.RemoveDuplicates, out *RemoveDuplicates, s conversion.Scope) error {
-	return autoConvert_api_RemoveDuplicates_To_v1alpha1_RemoveDuplicates(in, out, s)
-}
-
-func autoConvert_v1alpha1_StrategyParameters_To_api_StrategyParameters(in *StrategyParameters, out *api.StrategyParameters, s conversion.Scope) error {
-	out.NodeResourceUtilizationThresholds = (*api.NodeResourceUtilizationThresholds)(unsafe.Pointer(in.NodeResourceUtilizationThresholds))
-	out.NodeAffinityType = *(*[]string)(unsafe.Pointer(&in.NodeAffinityType))
-	out.PodsHavingTooManyRestarts = (*api.PodsHavingTooManyRestarts)(unsafe.Pointer(in.PodsHavingTooManyRestarts))
-	out.PodLifeTime = (*api.PodLifeTime)(unsafe.Pointer(in.PodLifeTime))
-	out.RemoveDuplicates = (*api.RemoveDuplicates)(unsafe.Pointer(in.RemoveDuplicates))
-	out.FailedPods = (*api.FailedPods)(unsafe.Pointer(in.FailedPods))
-	out.IncludeSoftConstraints = in.IncludeSoftConstraints
-	out.Namespaces = (*api.Namespaces)(unsafe.Pointer(in.Namespaces))
-	out.ThresholdPriority = (*int32)(unsafe.Pointer(in.ThresholdPriority))
-	out.ThresholdPriorityClassName = in.ThresholdPriorityClassName
-	out.LabelSelector = (*v1.LabelSelector)(unsafe.Pointer(in.LabelSelector))
-	out.NodeFit = in.NodeFit
-	return nil
-}
-
-// Convert_v1alpha1_StrategyParameters_To_api_StrategyParameters is an autogenerated conversion function.
-func Convert_v1alpha1_StrategyParameters_To_api_StrategyParameters(in *StrategyParameters, out *api.StrategyParameters, s conversion.Scope) error {
-	return autoConvert_v1alpha1_StrategyParameters_To_api_StrategyParameters(in, out, s)
-}
-
-func autoConvert_api_StrategyParameters_To_v1alpha1_StrategyParameters(in *api.StrategyParameters, out *StrategyParameters, s conversion.Scope) error {
-	out.NodeResourceUtilizationThresholds = (*NodeResourceUtilizationThresholds)(unsafe.Pointer(in.NodeResourceUtilizationThresholds))
-	out.NodeAffinityType = *(*[]string)(unsafe.Pointer(&in.NodeAffinityType))
-	out.PodsHavingTooManyRestarts = (*PodsHavingTooManyRestarts)(unsafe.Pointer(in.PodsHavingTooManyRestarts))
-	out.PodLifeTime = (*PodLifeTime)(unsafe.Pointer(in.PodLifeTime))
-	out.RemoveDuplicates = (*RemoveDuplicates)(unsafe.Pointer(in.RemoveDuplicates))
-	out.FailedPods = (*FailedPods)(unsafe.Pointer(in.FailedPods))
-	out.IncludeSoftConstraints = in.IncludeSoftConstraints
-	out.Namespaces = (*Namespaces)(unsafe.Pointer(in.Namespaces))
-	out.ThresholdPriority = (*int32)(unsafe.Pointer(in.ThresholdPriority))
-	out.ThresholdPriorityClassName = in.ThresholdPriorityClassName
-	out.LabelSelector = (*v1.LabelSelector)(unsafe.Pointer(in.LabelSelector))
-	out.NodeFit = in.NodeFit
-	return nil
-}
-
-// Convert_api_StrategyParameters_To_v1alpha1_StrategyParameters is an autogenerated conversion function.
-func Convert_api_StrategyParameters_To_v1alpha1_StrategyParameters(in *api.StrategyParameters, out *StrategyParameters, s conversion.Scope) error {
-	return autoConvert_api_StrategyParameters_To_v1alpha1_StrategyParameters(in, out, s)
-}

pkg/api/v1alpha1/zz_generated.deepcopy.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -41,6 +42,11 @@ func (in *DeschedulerPolicy) DeepCopyInto(out *DeschedulerPolicy) {
 		*out = new(string)
 		**out = **in
 	}
+	if in.EvictFailedBarePods != nil {
+		in, out := &in.EvictFailedBarePods, &out.EvictFailedBarePods
+		*out = new(bool)
+		**out = **in
+	}
 	if in.EvictLocalStoragePods != nil {
 		in, out := &in.EvictLocalStoragePods, &out.EvictLocalStoragePods
 		*out = new(bool)
@@ -58,7 +64,12 @@ func (in *DeschedulerPolicy) DeepCopyInto(out *DeschedulerPolicy) {
 	}
 	if in.MaxNoOfPodsToEvictPerNode != nil {
 		in, out := &in.MaxNoOfPodsToEvictPerNode, &out.MaxNoOfPodsToEvictPerNode
-		*out = new(int)
+		*out = new(uint)
+		**out = **in
+	}
+	if in.MaxNoOfPodsToEvictPerNamespace != nil {
+		in, out := &in.MaxNoOfPodsToEvictPerNamespace, &out.MaxNoOfPodsToEvictPerNamespace
+		*out = new(uint)
 		**out = **in
 	}
 	return
@@ -198,6 +209,11 @@ func (in *PodLifeTime) DeepCopyInto(out *PodLifeTime) {
 		*out = new(uint)
 		**out = **in
 	}
+	if in.States != nil {
+		in, out := &in.States, &out.States
+		*out = make([]string, len(*in))
+		copy(*out, *in)
+	}
 	if in.PodStatusPhases != nil {
 		in, out := &in.PodStatusPhases, &out.PodStatusPhases
 		*out = make([]string, len(*in))
@@ -345,6 +361,11 @@ func (in *StrategyParameters) DeepCopyInto(out *StrategyParameters) {
 		*out = new(v1.LabelSelector)
 		(*in).DeepCopyInto(*out)
 	}
+	if in.ExcludedTaints != nil {
+		in, out := &in.ExcludedTaints, &out.ExcludedTaints
+		*out = make([]string, len(*in))
+		copy(*out, *in)
+	}
 	return
 }
 

pkg/api/v1alpha1/zz_generated.defaults.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.

pkg/api/zz_generated.deepcopy.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -21,7 +22,6 @@ limitations under the License.
 package api
 
 import (
-	v1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	runtime "k8s.io/apimachinery/pkg/runtime"
 )
 
@@ -29,11 +29,11 @@ import (
 func (in *DeschedulerPolicy) DeepCopyInto(out *DeschedulerPolicy) {
 	*out = *in
 	out.TypeMeta = in.TypeMeta
-	if in.Strategies != nil {
-		in, out := &in.Strategies, &out.Strategies
-		*out = make(StrategyList, len(*in))
-		for key, val := range *in {
-			(*out)[key] = *val.DeepCopy()
+	if in.Profiles != nil {
+		in, out := &in.Profiles, &out.Profiles
+		*out = make([]Profile, len(*in))
+		for i := range *in {
+			(*in)[i].DeepCopyInto(&(*out)[i])
 		}
 	}
 	if in.NodeSelector != nil {
@@ -41,24 +41,14 @@ func (in *DeschedulerPolicy) DeepCopyInto(out *DeschedulerPolicy) {
 		*out = new(string)
 		**out = **in
 	}
-	if in.EvictLocalStoragePods != nil {
-		in, out := &in.EvictLocalStoragePods, &out.EvictLocalStoragePods
-		*out = new(bool)
-		**out = **in
-	}
-	if in.EvictSystemCriticalPods != nil {
-		in, out := &in.EvictSystemCriticalPods, &out.EvictSystemCriticalPods
-		*out = new(bool)
-		**out = **in
-	}
-	if in.IgnorePVCPods != nil {
-		in, out := &in.IgnorePVCPods, &out.IgnorePVCPods
-		*out = new(bool)
-		**out = **in
-	}
 	if in.MaxNoOfPodsToEvictPerNode != nil {
 		in, out := &in.MaxNoOfPodsToEvictPerNode, &out.MaxNoOfPodsToEvictPerNode
-		*out = new(int)
+		*out = new(uint)
+		**out = **in
+	}
+	if in.MaxNoOfPodsToEvictPerNamespace != nil {
+		in, out := &in.MaxNoOfPodsToEvictPerNamespace, &out.MaxNoOfPodsToEvictPerNamespace
+		*out = new(uint)
 		**out = **in
 	}
 	return
@@ -82,58 +72,6 @@ func (in *DeschedulerPolicy) DeepCopyObject() runtime.Object {
 	return nil
 }
 
-// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *DeschedulerStrategy) DeepCopyInto(out *DeschedulerStrategy) {
-	*out = *in
-	if in.Params != nil {
-		in, out := &in.Params, &out.Params
-		*out = new(StrategyParameters)
-		(*in).DeepCopyInto(*out)
-	}
-	return
-}
-
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new DeschedulerStrategy.
-func (in *DeschedulerStrategy) DeepCopy() *DeschedulerStrategy {
-	if in == nil {
-		return nil
-	}
-	out := new(DeschedulerStrategy)
-	in.DeepCopyInto(out)
-	return out
-}
-
-// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *FailedPods) DeepCopyInto(out *FailedPods) {
-	*out = *in
-	if in.ExcludeOwnerKinds != nil {
-		in, out := &in.ExcludeOwnerKinds, &out.ExcludeOwnerKinds
-		*out = make([]string, len(*in))
-		copy(*out, *in)
-	}
-	if in.MinPodLifetimeSeconds != nil {
-		in, out := &in.MinPodLifetimeSeconds, &out.MinPodLifetimeSeconds
-		*out = new(uint)
-		**out = **in
-	}
-	if in.Reasons != nil {
-		in, out := &in.Reasons, &out.Reasons
-		*out = make([]string, len(*in))
-		copy(*out, *in)
-	}
-	return
-}
-
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new FailedPods.
-func (in *FailedPods) DeepCopy() *FailedPods {
-	if in == nil {
-		return nil
-	}
-	out := new(FailedPods)
-	in.DeepCopyInto(out)
-	return out
-}
-
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
 func (in *Namespaces) DeepCopyInto(out *Namespaces) {
 	*out = *in
@@ -161,94 +99,114 @@ func (in *Namespaces) DeepCopy() *Namespaces {
 }
 
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *NodeResourceUtilizationThresholds) DeepCopyInto(out *NodeResourceUtilizationThresholds) {
+func (in *PluginConfig) DeepCopyInto(out *PluginConfig) {
 	*out = *in
-	if in.Thresholds != nil {
-		in, out := &in.Thresholds, &out.Thresholds
-		*out = make(ResourceThresholds, len(*in))
-		for key, val := range *in {
-			(*out)[key] = val
-		}
-	}
-	if in.TargetThresholds != nil {
-		in, out := &in.TargetThresholds, &out.TargetThresholds
-		*out = make(ResourceThresholds, len(*in))
-		for key, val := range *in {
-			(*out)[key] = val
-		}
+	if in.Args != nil {
+		out.Args = in.Args.DeepCopyObject()
 	}
 	return
 }
 
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new NodeResourceUtilizationThresholds.
-func (in *NodeResourceUtilizationThresholds) DeepCopy() *NodeResourceUtilizationThresholds {
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PluginConfig.
+func (in *PluginConfig) DeepCopy() *PluginConfig {
 	if in == nil {
 		return nil
 	}
-	out := new(NodeResourceUtilizationThresholds)
+	out := new(PluginConfig)
 	in.DeepCopyInto(out)
 	return out
 }
 
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *PodLifeTime) DeepCopyInto(out *PodLifeTime) {
+func (in *PluginSet) DeepCopyInto(out *PluginSet) {
 	*out = *in
-	if in.MaxPodLifeTimeSeconds != nil {
-		in, out := &in.MaxPodLifeTimeSeconds, &out.MaxPodLifeTimeSeconds
-		*out = new(uint)
+	if in.Enabled != nil {
+		in, out := &in.Enabled, &out.Enabled
+		*out = make([]string, len(*in))
+		copy(*out, *in)
+	}
+	if in.Disabled != nil {
+		in, out := &in.Disabled, &out.Disabled
+		*out = make([]string, len(*in))
+		copy(*out, *in)
+	}
+	return
+}
+
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PluginSet.
+func (in *PluginSet) DeepCopy() *PluginSet {
+	if in == nil {
+		return nil
+	}
+	out := new(PluginSet)
+	in.DeepCopyInto(out)
+	return out
+}
+
+// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
+func (in *Plugins) DeepCopyInto(out *Plugins) {
+	*out = *in
+	in.PreSort.DeepCopyInto(&out.PreSort)
+	in.Sort.DeepCopyInto(&out.Sort)
+	in.Deschedule.DeepCopyInto(&out.Deschedule)
+	in.Balance.DeepCopyInto(&out.Balance)
+	in.Evict.DeepCopyInto(&out.Evict)
+	in.Filter.DeepCopyInto(&out.Filter)
+	in.PreEvictionFilter.DeepCopyInto(&out.PreEvictionFilter)
+	return
+}
+
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new Plugins.
+func (in *Plugins) DeepCopy() *Plugins {
+	if in == nil {
+		return nil
+	}
+	out := new(Plugins)
+	in.DeepCopyInto(out)
+	return out
+}
+
+// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
+func (in *PriorityThreshold) DeepCopyInto(out *PriorityThreshold) {
+	*out = *in
+	if in.Value != nil {
+		in, out := &in.Value, &out.Value
+		*out = new(int32)
 		**out = **in
 	}
-	if in.PodStatusPhases != nil {
-		in, out := &in.PodStatusPhases, &out.PodStatusPhases
-		*out = make([]string, len(*in))
-		copy(*out, *in)
-	}
 	return
 }
 
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PodLifeTime.
-func (in *PodLifeTime) DeepCopy() *PodLifeTime {
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PriorityThreshold.
+func (in *PriorityThreshold) DeepCopy() *PriorityThreshold {
 	if in == nil {
 		return nil
 	}
-	out := new(PodLifeTime)
+	out := new(PriorityThreshold)
 	in.DeepCopyInto(out)
 	return out
 }
 
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *PodsHavingTooManyRestarts) DeepCopyInto(out *PodsHavingTooManyRestarts) {
+func (in *Profile) DeepCopyInto(out *Profile) {
 	*out = *in
+	if in.PluginConfigs != nil {
+		in, out := &in.PluginConfigs, &out.PluginConfigs
+		*out = make([]PluginConfig, len(*in))
+		for i := range *in {
+			(*in)[i].DeepCopyInto(&(*out)[i])
+		}
+	}
+	in.Plugins.DeepCopyInto(&out.Plugins)
 	return
 }
 
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PodsHavingTooManyRestarts.
-func (in *PodsHavingTooManyRestarts) DeepCopy() *PodsHavingTooManyRestarts {
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new Profile.
+func (in *Profile) DeepCopy() *Profile {
 	if in == nil {
 		return nil
 	}
-	out := new(PodsHavingTooManyRestarts)
-	in.DeepCopyInto(out)
-	return out
-}
-
-// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *RemoveDuplicates) DeepCopyInto(out *RemoveDuplicates) {
-	*out = *in
-	if in.ExcludeOwnerKinds != nil {
-		in, out := &in.ExcludeOwnerKinds, &out.ExcludeOwnerKinds
-		*out = make([]string, len(*in))
-		copy(*out, *in)
-	}
-	return
-}
-
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new RemoveDuplicates.
-func (in *RemoveDuplicates) DeepCopy() *RemoveDuplicates {
-	if in == nil {
-		return nil
-	}
-	out := new(RemoveDuplicates)
+	out := new(Profile)
 	in.DeepCopyInto(out)
 	return out
 }
@@ -274,86 +232,3 @@ func (in ResourceThresholds) DeepCopy() ResourceThresholds {
 	in.DeepCopyInto(out)
 	return *out
 }
-
-// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in StrategyList) DeepCopyInto(out *StrategyList) {
-	{
-		in := &in
-		*out = make(StrategyList, len(*in))
-		for key, val := range *in {
-			(*out)[key] = *val.DeepCopy()
-		}
-		return
-	}
-}
-
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new StrategyList.
-func (in StrategyList) DeepCopy() StrategyList {
-	if in == nil {
-		return nil
-	}
-	out := new(StrategyList)
-	in.DeepCopyInto(out)
-	return *out
-}
-
-// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
-func (in *StrategyParameters) DeepCopyInto(out *StrategyParameters) {
-	*out = *in
-	if in.NodeResourceUtilizationThresholds != nil {
-		in, out := &in.NodeResourceUtilizationThresholds, &out.NodeResourceUtilizationThresholds
-		*out = new(NodeResourceUtilizationThresholds)
-		(*in).DeepCopyInto(*out)
-	}
-	if in.NodeAffinityType != nil {
-		in, out := &in.NodeAffinityType, &out.NodeAffinityType
-		*out = make([]string, len(*in))
-		copy(*out, *in)
-	}
-	if in.PodsHavingTooManyRestarts != nil {
-		in, out := &in.PodsHavingTooManyRestarts, &out.PodsHavingTooManyRestarts
-		*out = new(PodsHavingTooManyRestarts)
-		**out = **in
-	}
-	if in.PodLifeTime != nil {
-		in, out := &in.PodLifeTime, &out.PodLifeTime
-		*out = new(PodLifeTime)
-		(*in).DeepCopyInto(*out)
-	}
-	if in.RemoveDuplicates != nil {
-		in, out := &in.RemoveDuplicates, &out.RemoveDuplicates
-		*out = new(RemoveDuplicates)
-		(*in).DeepCopyInto(*out)
-	}
-	if in.FailedPods != nil {
-		in, out := &in.FailedPods, &out.FailedPods
-		*out = new(FailedPods)
-		(*in).DeepCopyInto(*out)
-	}
-	if in.Namespaces != nil {
-		in, out := &in.Namespaces, &out.Namespaces
-		*out = new(Namespaces)
-		(*in).DeepCopyInto(*out)
-	}
-	if in.ThresholdPriority != nil {
-		in, out := &in.ThresholdPriority, &out.ThresholdPriority
-		*out = new(int32)
-		**out = **in
-	}
-	if in.LabelSelector != nil {
-		in, out := &in.LabelSelector, &out.LabelSelector
-		*out = new(v1.LabelSelector)
-		(*in).DeepCopyInto(*out)
-	}
-	return
-}
-
-// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new StrategyParameters.
-func (in *StrategyParameters) DeepCopy() *StrategyParameters {
-	if in == nil {
-		return nil
-	}
-	out := new(StrategyParameters)
-	in.DeepCopyInto(out)
-	return out
-}

pkg/apis/componentconfig/types.go

@@ -21,6 +21,7 @@ import (
 
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	componentbaseconfig "k8s.io/component-base/config"
+	registry "k8s.io/component-base/logs/api/v1"
 )
 
 // +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
@@ -53,7 +54,10 @@ type DeschedulerConfiguration struct {
 	// IgnorePVCPods sets whether PVC pods should be allowed to be evicted
 	IgnorePVCPods bool
 
+	// LeaderElection starts Deployment using leader election loop
+	LeaderElection componentbaseconfig.LeaderElectionConfiguration
+
 	// Logging specifies the options of logging.
-	// Refer [Logs Options](https://github.com/kubernetes/component-base/blob/master/logs/options.go) for more information.
-	Logging componentbaseconfig.LoggingConfiguration
+	// Refer [Logs Options](https://github.com/kubernetes/component-base/blob/master/logs/api/v1/options.go) for more information.
+	Logging registry.LoggingConfiguration
 }

pkg/apis/componentconfig/v1alpha1/register.go

@@ -28,8 +28,10 @@ var (
 )
 
 // GroupName is the group name use in this package
-const GroupName = "deschedulercomponentconfig"
-const GroupVersion = "v1alpha1"
+const (
+	GroupName    = "deschedulercomponentconfig"
+	GroupVersion = "v1alpha1"
+)
 
 // SchemeGroupVersion is group version used to register these objects
 var SchemeGroupVersion = schema.GroupVersion{Group: GroupName, Version: GroupVersion}

pkg/apis/componentconfig/v1alpha1/types.go

@@ -17,10 +17,11 @@ limitations under the License.
 package v1alpha1
 
 import (
-	componentbaseconfig "k8s.io/component-base/config"
 	"time"
 
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	componentbaseconfig "k8s.io/component-base/config"
+	registry "k8s.io/component-base/logs/api/v1"
 )
 
 // +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object
@@ -53,7 +54,10 @@ type DeschedulerConfiguration struct {
 	// IgnorePVCPods sets whether PVC pods should be allowed to be evicted
 	IgnorePVCPods bool `json:"ignorePvcPods,omitempty"`
 
+	// LeaderElection starts Deployment using leader election loop
+	LeaderElection componentbaseconfig.LeaderElectionConfiguration `json:"leaderElection,omitempty"`
+
 	// Logging specifies the options of logging.
-	// Refer [Logs Options](https://github.com/kubernetes/component-base/blob/master/logs/options.go) for more information.
-	Logging componentbaseconfig.LoggingConfiguration `json:"logging,omitempty"`
+	// Refer [Logs Options](https://github.com/kubernetes/component-base/blob/master/logs/api/v1/options.go) for more information.
+	Logging registry.LoggingConfiguration `json:"logging,omitempty"`
 }

pkg/apis/componentconfig/v1alpha1/zz_generated.conversion.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -57,6 +58,7 @@ func autoConvert_v1alpha1_DeschedulerConfiguration_To_componentconfig_Deschedule
 	out.MaxNoOfPodsToEvictPerNode = in.MaxNoOfPodsToEvictPerNode
 	out.EvictLocalStoragePods = in.EvictLocalStoragePods
 	out.IgnorePVCPods = in.IgnorePVCPods
+	out.LeaderElection = in.LeaderElection
 	out.Logging = in.Logging
 	return nil
 }
@@ -75,6 +77,7 @@ func autoConvert_componentconfig_DeschedulerConfiguration_To_v1alpha1_Deschedule
 	out.MaxNoOfPodsToEvictPerNode = in.MaxNoOfPodsToEvictPerNode
 	out.EvictLocalStoragePods = in.EvictLocalStoragePods
 	out.IgnorePVCPods = in.IgnorePVCPods
+	out.LeaderElection = in.LeaderElection
 	out.Logging = in.Logging
 	return nil
 }

pkg/apis/componentconfig/v1alpha1/zz_generated.deepcopy.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -28,7 +29,8 @@ import (
 func (in *DeschedulerConfiguration) DeepCopyInto(out *DeschedulerConfiguration) {
 	*out = *in
 	out.TypeMeta = in.TypeMeta
-	out.Logging = in.Logging
+	out.LeaderElection = in.LeaderElection
+	in.Logging.DeepCopyInto(&out.Logging)
 	return
 }
 

pkg/apis/componentconfig/v1alpha1/zz_generated.defaults.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.

pkg/apis/componentconfig/zz_generated.deepcopy.go

@@ -1,7 +1,8 @@
+//go:build !ignore_autogenerated
 // +build !ignore_autogenerated
 
 /*
-Copyright 2021 The Kubernetes Authors.
+Copyright 2023 The Kubernetes Authors.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -28,7 +29,8 @@ import (
 func (in *DeschedulerConfiguration) DeepCopyInto(out *DeschedulerConfiguration) {
 	*out = *in
 	out.TypeMeta = in.TypeMeta
-	out.Logging = in.Logging
+	out.LeaderElection = in.LeaderElection
+	in.Logging.DeepCopyInto(&out.Logging)
 	return
 }
 

pkg/descheduler/client/client.go

@@ -26,7 +26,7 @@ import (
 	"k8s.io/client-go/tools/clientcmd"
 )
 
-func CreateClient(kubeconfig string) (clientset.Interface, error) {
+func CreateClient(kubeconfig, userAgt string) (clientset.Interface, error) {
 	var cfg *rest.Config
 	if len(kubeconfig) != 0 {
 		master, err := GetMasterFromKubeconfig(kubeconfig)
@@ -47,7 +47,11 @@ func CreateClient(kubeconfig string) (clientset.Interface, error) {
 		}
 	}
 
-	return clientset.NewForConfig(cfg)
+	if len(userAgt) != 0 {
+		return clientset.NewForConfig(rest.AddUserAgent(cfg, userAgt))
+	} else {
+		return clientset.NewForConfig(cfg)
+	}
 }
 
 func GetMasterFromKubeconfig(filename string) (string, error) {

pkg/descheduler/descheduler.go

@@ -19,14 +19,22 @@ package descheduler
 import (
 	"context"
 	"fmt"
-	"sigs.k8s.io/descheduler/pkg/descheduler/strategies/nodeutilization"
 
-	v1 "k8s.io/api/core/v1"
-	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog/v2"
 
+	v1 "k8s.io/api/core/v1"
+	policy "k8s.io/api/policy/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/labels"
+	"k8s.io/apimachinery/pkg/runtime"
 	"k8s.io/apimachinery/pkg/util/wait"
 	"k8s.io/client-go/informers"
+	clientset "k8s.io/client-go/kubernetes"
+	fakeclientset "k8s.io/client-go/kubernetes/fake"
+	listersv1 "k8s.io/client-go/listers/core/v1"
+	schedulingv1 "k8s.io/client-go/listers/scheduling/v1"
+	core "k8s.io/client-go/testing"
+
 	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"sigs.k8s.io/descheduler/metrics"
 	"sigs.k8s.io/descheduler/pkg/api"
@@ -34,20 +42,24 @@ import (
 	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
 	eutils "sigs.k8s.io/descheduler/pkg/descheduler/evictions/utils"
 	nodeutil "sigs.k8s.io/descheduler/pkg/descheduler/node"
-	"sigs.k8s.io/descheduler/pkg/descheduler/strategies"
+	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
+	"sigs.k8s.io/descheduler/pkg/framework"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/defaultevictor"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/pluginbuilder"
+	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
-func Run(rs *options.DeschedulerServer) error {
+func Run(ctx context.Context, rs *options.DeschedulerServer) error {
 	metrics.Register()
 
-	ctx := context.Background()
-	rsclient, err := client.CreateClient(rs.KubeconfigFile)
+	rsclient, eventClient, err := createClients(rs.KubeconfigFile)
 	if err != nil {
 		return err
 	}
 	rs.Client = rsclient
+	rs.EventClient = eventClient
 
-	deschedulerPolicy, err := LoadPolicyConfig(rs.PolicyConfigFile)
+	deschedulerPolicy, err := LoadPolicyConfig(rs.PolicyConfigFile, rs.Client, pluginbuilder.PluginRegistry)
 	if err != nil {
 		return err
 	}
@@ -60,92 +72,336 @@ func Run(rs *options.DeschedulerServer) error {
 		return err
 	}
 
-	stopChannel := make(chan struct{})
-	return RunDeschedulerStrategies(ctx, rs, deschedulerPolicy, evictionPolicyGroupVersion, stopChannel)
-}
-
-type strategyFunction func(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, podEvictor *evictions.PodEvictor)
-
-func RunDeschedulerStrategies(ctx context.Context, rs *options.DeschedulerServer, deschedulerPolicy *api.DeschedulerPolicy, evictionPolicyGroupVersion string, stopChannel chan struct{}) error {
-	sharedInformerFactory := informers.NewSharedInformerFactory(rs.Client, 0)
-	nodeInformer := sharedInformerFactory.Core().V1().Nodes()
-
-	sharedInformerFactory.Start(stopChannel)
-	sharedInformerFactory.WaitForCacheSync(stopChannel)
-
-	strategyFuncs := map[api.StrategyName]strategyFunction{
-		"RemoveDuplicates":                            strategies.RemoveDuplicatePods,
-		"LowNodeUtilization":                          nodeutilization.LowNodeUtilization,
-		"HighNodeUtilization":                         nodeutilization.HighNodeUtilization,
-		"RemovePodsViolatingInterPodAntiAffinity":     strategies.RemovePodsViolatingInterPodAntiAffinity,
-		"RemovePodsViolatingNodeAffinity":             strategies.RemovePodsViolatingNodeAffinity,
-		"RemovePodsViolatingNodeTaints":               strategies.RemovePodsViolatingNodeTaints,
-		"RemovePodsHavingTooManyRestarts":             strategies.RemovePodsHavingTooManyRestarts,
-		"PodLifeTime":                                 strategies.PodLifeTime,
-		"RemovePodsViolatingTopologySpreadConstraint": strategies.RemovePodsViolatingTopologySpreadConstraint,
-		"RemoveFailedPods":                            strategies.RemoveFailedPods,
+	runFn := func() error {
+		return RunDeschedulerStrategies(ctx, rs, deschedulerPolicy, evictionPolicyGroupVersion)
 	}
 
-	nodeSelector := rs.NodeSelector
+	if rs.LeaderElection.LeaderElect && rs.DeschedulingInterval.Seconds() == 0 {
+		return fmt.Errorf("leaderElection must be used with deschedulingInterval")
+	}
+
+	if rs.LeaderElection.LeaderElect && rs.DryRun {
+		klog.V(1).InfoS("Warning: DryRun is set to True. You need to disable it to use Leader Election.")
+	}
+
+	if rs.LeaderElection.LeaderElect && !rs.DryRun {
+		if err := NewLeaderElection(runFn, rsclient, &rs.LeaderElection, ctx); err != nil {
+			return fmt.Errorf("leaderElection: %w", err)
+		}
+		return nil
+	}
+
+	return runFn()
+}
+
+func cachedClient(
+	realClient clientset.Interface,
+	podLister listersv1.PodLister,
+	nodeLister listersv1.NodeLister,
+	namespaceLister listersv1.NamespaceLister,
+	priorityClassLister schedulingv1.PriorityClassLister,
+) (clientset.Interface, error) {
+	fakeClient := fakeclientset.NewSimpleClientset()
+	// simulate a pod eviction by deleting a pod
+	fakeClient.PrependReactor("create", "pods", func(action core.Action) (bool, runtime.Object, error) {
+		if action.GetSubresource() == "eviction" {
+			createAct, matched := action.(core.CreateActionImpl)
+			if !matched {
+				return false, nil, fmt.Errorf("unable to convert action to core.CreateActionImpl")
+			}
+			eviction, matched := createAct.Object.(*policy.Eviction)
+			if !matched {
+				return false, nil, fmt.Errorf("unable to convert action object into *policy.Eviction")
+			}
+			if err := fakeClient.Tracker().Delete(action.GetResource(), eviction.GetNamespace(), eviction.GetName()); err != nil {
+				return false, nil, fmt.Errorf("unable to delete pod %v/%v: %v", eviction.GetNamespace(), eviction.GetName(), err)
+			}
+			return true, nil, nil
+		}
+		// fallback to the default reactor
+		return false, nil, nil
+	})
+
+	klog.V(3).Infof("Pulling resources for the cached client from the cluster")
+	pods, err := podLister.List(labels.Everything())
+	if err != nil {
+		return nil, fmt.Errorf("unable to list pods: %v", err)
+	}
+
+	for _, item := range pods {
+		if _, err := fakeClient.CoreV1().Pods(item.Namespace).Create(context.TODO(), item, metav1.CreateOptions{}); err != nil {
+			return nil, fmt.Errorf("unable to copy pod: %v", err)
+		}
+	}
+
+	nodes, err := nodeLister.List(labels.Everything())
+	if err != nil {
+		return nil, fmt.Errorf("unable to list nodes: %v", err)
+	}
+
+	for _, item := range nodes {
+		if _, err := fakeClient.CoreV1().Nodes().Create(context.TODO(), item, metav1.CreateOptions{}); err != nil {
+			return nil, fmt.Errorf("unable to copy node: %v", err)
+		}
+	}
+
+	namespaces, err := namespaceLister.List(labels.Everything())
+	if err != nil {
+		return nil, fmt.Errorf("unable to list namespaces: %v", err)
+	}
+
+	for _, item := range namespaces {
+		if _, err := fakeClient.CoreV1().Namespaces().Create(context.TODO(), item, metav1.CreateOptions{}); err != nil {
+			return nil, fmt.Errorf("unable to copy node: %v", err)
+		}
+	}
+
+	priorityClasses, err := priorityClassLister.List(labels.Everything())
+	if err != nil {
+		return nil, fmt.Errorf("unable to list priorityclasses: %v", err)
+	}
+
+	for _, item := range priorityClasses {
+		if _, err := fakeClient.SchedulingV1().PriorityClasses().Create(context.TODO(), item, metav1.CreateOptions{}); err != nil {
+			return nil, fmt.Errorf("unable to copy priorityclass: %v", err)
+		}
+	}
+
+	return fakeClient, nil
+}
+
+// evictorImpl implements the Evictor interface so plugins
+// can evict a pod without importing a specific pod evictor
+type evictorImpl struct {
+	podEvictor    *evictions.PodEvictor
+	evictorFilter framework.EvictorPlugin
+}
+
+var _ framework.Evictor = &evictorImpl{}
+
+// Filter checks if a pod can be evicted
+func (ei *evictorImpl) Filter(pod *v1.Pod) bool {
+	return ei.evictorFilter.Filter(pod)
+}
+
+// PreEvictionFilter checks if pod can be evicted right before eviction
+func (ei *evictorImpl) PreEvictionFilter(pod *v1.Pod) bool {
+	return ei.evictorFilter.PreEvictionFilter(pod)
+}
+
+// Evict evicts a pod (no pre-check performed)
+func (ei *evictorImpl) Evict(ctx context.Context, pod *v1.Pod, opts evictions.EvictOptions) bool {
+	return ei.podEvictor.EvictPod(ctx, pod, opts)
+}
+
+func (ei *evictorImpl) NodeLimitExceeded(node *v1.Node) bool {
+	return ei.podEvictor.NodeLimitExceeded(node)
+}
+
+// handleImpl implements the framework handle which gets passed to plugins
+type handleImpl struct {
+	clientSet                 clientset.Interface
+	getPodsAssignedToNodeFunc podutil.GetPodsAssignedToNodeFunc
+	sharedInformerFactory     informers.SharedInformerFactory
+	evictor                   *evictorImpl
+}
+
+var _ framework.Handle = &handleImpl{}
+
+// ClientSet retrieves kube client set
+func (hi *handleImpl) ClientSet() clientset.Interface {
+	return hi.clientSet
+}
+
+// GetPodsAssignedToNodeFunc retrieves GetPodsAssignedToNodeFunc implementation
+func (hi *handleImpl) GetPodsAssignedToNodeFunc() podutil.GetPodsAssignedToNodeFunc {
+	return hi.getPodsAssignedToNodeFunc
+}
+
+// SharedInformerFactory retrieves shared informer factory
+func (hi *handleImpl) SharedInformerFactory() informers.SharedInformerFactory {
+	return hi.sharedInformerFactory
+}
+
+// Evictor retrieves evictor so plugins can filter and evict pods
+func (hi *handleImpl) Evictor() framework.Evictor {
+	return hi.evictor
+}
+
+func RunDeschedulerStrategies(ctx context.Context, rs *options.DeschedulerServer, deschedulerPolicy *api.DeschedulerPolicy, evictionPolicyGroupVersion string) error {
+	sharedInformerFactory := informers.NewSharedInformerFactory(rs.Client, 0)
+	podInformer := sharedInformerFactory.Core().V1().Pods().Informer()
+	podLister := sharedInformerFactory.Core().V1().Pods().Lister()
+	nodeLister := sharedInformerFactory.Core().V1().Nodes().Lister()
+	namespaceLister := sharedInformerFactory.Core().V1().Namespaces().Lister()
+	priorityClassLister := sharedInformerFactory.Scheduling().V1().PriorityClasses().Lister()
+
+	ctx, cancel := context.WithCancel(ctx)
+	defer cancel()
+
+	getPodsAssignedToNode, err := podutil.BuildGetPodsAssignedToNodeFunc(podInformer)
+	if err != nil {
+		return fmt.Errorf("build get pods assigned to node function error: %v", err)
+	}
+
+	sharedInformerFactory.Start(ctx.Done())
+	sharedInformerFactory.WaitForCacheSync(ctx.Done())
+
+	var nodeSelector string
 	if deschedulerPolicy.NodeSelector != nil {
 		nodeSelector = *deschedulerPolicy.NodeSelector
 	}
 
-	evictLocalStoragePods := rs.EvictLocalStoragePods
-	if deschedulerPolicy.EvictLocalStoragePods != nil {
-		evictLocalStoragePods = *deschedulerPolicy.EvictLocalStoragePods
+	var eventClient clientset.Interface
+	if rs.DryRun {
+		eventClient = fakeclientset.NewSimpleClientset()
+	} else {
+		eventClient = rs.Client
 	}
 
-	evictSystemCriticalPods := false
-	if deschedulerPolicy.EvictSystemCriticalPods != nil {
-		evictSystemCriticalPods = *deschedulerPolicy.EvictSystemCriticalPods
-		if evictSystemCriticalPods {
-			klog.V(1).InfoS("Warning: EvictSystemCriticalPods is set to True. This could cause eviction of Kubernetes system pods.")
-		}
-	}
+	eventBroadcaster, eventRecorder := utils.GetRecorderAndBroadcaster(ctx, eventClient)
+	defer eventBroadcaster.Shutdown()
 
-	ignorePvcPods := false
-	if deschedulerPolicy.IgnorePVCPods != nil {
-		ignorePvcPods = *deschedulerPolicy.IgnorePVCPods
-	}
-
-	maxNoOfPodsToEvictPerNode := rs.MaxNoOfPodsToEvictPerNode
-	if deschedulerPolicy.MaxNoOfPodsToEvictPerNode != nil {
-		maxNoOfPodsToEvictPerNode = *deschedulerPolicy.MaxNoOfPodsToEvictPerNode
-	}
-
-	wait.Until(func() {
-		nodes, err := nodeutil.ReadyNodes(ctx, rs.Client, nodeInformer, nodeSelector)
+	wait.NonSlidingUntil(func() {
+		nodes, err := nodeutil.ReadyNodes(ctx, rs.Client, nodeLister, nodeSelector)
 		if err != nil {
 			klog.V(1).InfoS("Unable to get ready nodes", "err", err)
-			close(stopChannel)
+			cancel()
 			return
 		}
 
 		if len(nodes) <= 1 {
 			klog.V(1).InfoS("The cluster size is 0 or 1 meaning eviction causes service disruption or degradation. So aborting..")
-			close(stopChannel)
+			cancel()
 			return
 		}
 
+		var podEvictorClient clientset.Interface
+		// When the dry mode is enable, collect all the relevant objects (mostly pods) under a fake client.
+		// So when evicting pods while running multiple strategies in a row have the cummulative effect
+		// as is when evicting pods for real.
+		if rs.DryRun {
+			klog.V(3).Infof("Building a cached client from the cluster for the dry run")
+			// Create a new cache so we start from scratch without any leftovers
+			fakeClient, err := cachedClient(rs.Client, podLister, nodeLister, namespaceLister, priorityClassLister)
+			if err != nil {
+				klog.Error(err)
+				return
+			}
+
+			fakeSharedInformerFactory := informers.NewSharedInformerFactory(fakeClient, 0)
+			getPodsAssignedToNode, err = podutil.BuildGetPodsAssignedToNodeFunc(fakeSharedInformerFactory.Core().V1().Pods().Informer())
+			if err != nil {
+				klog.Errorf("build get pods assigned to node function error: %v", err)
+				return
+			}
+
+			fakeCtx, cncl := context.WithCancel(context.TODO())
+			defer cncl()
+			fakeSharedInformerFactory.Start(fakeCtx.Done())
+			fakeSharedInformerFactory.WaitForCacheSync(fakeCtx.Done())
+
+			podEvictorClient = fakeClient
+		} else {
+			podEvictorClient = rs.Client
+		}
+
+		klog.V(3).Infof("Building a pod evictor")
 		podEvictor := evictions.NewPodEvictor(
-			rs.Client,
+			podEvictorClient,
 			evictionPolicyGroupVersion,
 			rs.DryRun,
-			maxNoOfPodsToEvictPerNode,
+			deschedulerPolicy.MaxNoOfPodsToEvictPerNode,
+			deschedulerPolicy.MaxNoOfPodsToEvictPerNamespace,
 			nodes,
-			evictLocalStoragePods,
-			evictSystemCriticalPods,
-			ignorePvcPods,
+			!rs.DisableMetrics,
+			eventRecorder,
 		)
 
-		for name, strategy := range deschedulerPolicy.Strategies {
-			if f, ok := strategyFuncs[name]; ok {
-				if strategy.Enabled {
-					f(ctx, rs.Client, strategy, nodes, podEvictor)
+		var enabledDeschedulePlugins []framework.DeschedulePlugin
+		var enabledBalancePlugins []framework.BalancePlugin
+
+		// Build plugins
+		for _, profile := range deschedulerPolicy.Profiles {
+			pc := getPluginConfig(defaultevictor.PluginName, profile.PluginConfigs)
+			if pc == nil {
+				klog.ErrorS(fmt.Errorf("unable to get plugin config"), "skipping plugin", "plugin", defaultevictor.PluginName, "profile", profile.Name)
+				continue
+			}
+			evictorFilter, err := defaultevictor.New(
+				pc.Args,
+				&handleImpl{
+					clientSet:                 rs.Client,
+					getPodsAssignedToNodeFunc: getPodsAssignedToNode,
+					sharedInformerFactory:     sharedInformerFactory,
+				},
+			)
+			if err != nil {
+				klog.ErrorS(fmt.Errorf("unable to construct a plugin"), "skipping plugin", "plugin", defaultevictor.PluginName)
+				continue
+			}
+			handle := &handleImpl{
+				clientSet:                 rs.Client,
+				getPodsAssignedToNodeFunc: getPodsAssignedToNode,
+				sharedInformerFactory:     sharedInformerFactory,
+				evictor: &evictorImpl{
+					podEvictor:    podEvictor,
+					evictorFilter: evictorFilter.(framework.EvictorPlugin),
+				},
+			}
+			// Assuming only a list of enabled extension points.
+			// Later, when a default list of plugins and their extension points is established,
+			// compute the list of enabled extension points as (DefaultEnabled + Enabled - Disabled)
+			for _, plugin := range append(profile.Plugins.Deschedule.Enabled, profile.Plugins.Balance.Enabled...) {
+				pc := getPluginConfig(plugin, profile.PluginConfigs)
+				if pc == nil {
+					klog.ErrorS(fmt.Errorf("unable to get plugin config"), "skipping plugin", "plugin", plugin)
+					continue
 				}
-			} else {
-				klog.ErrorS(fmt.Errorf("unknown strategy name"), "skipping strategy", "strategy", name)
+				registryPlugin, ok := pluginbuilder.PluginRegistry[plugin]
+				pgFnc := registryPlugin.PluginBuilder
+				if !ok {
+					klog.ErrorS(fmt.Errorf("unable to find plugin in the pluginsMap"), "skipping plugin", "plugin", plugin)
+				}
+				pg, err := pgFnc(pc.Args, handle)
+				if err != nil {
+					klog.ErrorS(err, "unable to initialize a plugin", "pluginName", plugin)
+				}
+				if pg != nil {
+					switch v := pg.(type) {
+					case framework.DeschedulePlugin:
+						enabledDeschedulePlugins = append(enabledDeschedulePlugins, v)
+					case framework.BalancePlugin:
+						enabledBalancePlugins = append(enabledBalancePlugins, v)
+					default:
+						klog.ErrorS(fmt.Errorf("unknown plugin extension point"), "skipping plugin", "plugin", plugin)
+					}
+				}
+			}
+		}
+
+		// Execute extension points
+		for _, pg := range enabledDeschedulePlugins {
+			// TODO: strategyName should be accessible from within the strategy using a framework
+			// handle or function which the Evictor has access to. For migration/in-progress framework
+			// work, we are currently passing this via context. To be removed
+			// (See discussion thread https://github.com/kubernetes-sigs/descheduler/pull/885#discussion_r919962292)
+			childCtx := context.WithValue(ctx, "strategyName", pg.Name())
+			status := pg.Deschedule(childCtx, nodes)
+			if status != nil && status.Err != nil {
+				klog.ErrorS(status.Err, "plugin finished with error", "pluginName", pg.Name())
+			}
+		}
+
+		for _, pg := range enabledBalancePlugins {
+			// TODO: strategyName should be accessible from within the strategy using a framework
+			// handle or function which the Evictor has access to. For migration/in-progress framework
+			// work, we are currently passing this via context. To be removed
+			// (See discussion thread https://github.com/kubernetes-sigs/descheduler/pull/885#discussion_r919962292)
+			childCtx := context.WithValue(ctx, "strategyName", pg.Name())
+			status := pg.Balance(childCtx, nodes)
+			if status != nil && status.Err != nil {
+				klog.ErrorS(status.Err, "plugin finished with error", "pluginName", pg.Name())
 			}
 		}
 
@@ -153,9 +409,32 @@ func RunDeschedulerStrategies(ctx context.Context, rs *options.DeschedulerServer
 
 		// If there was no interval specified, send a signal to the stopChannel to end the wait.Until loop after 1 iteration
 		if rs.DeschedulingInterval.Seconds() == 0 {
-			close(stopChannel)
+			cancel()
 		}
-	}, rs.DeschedulingInterval, stopChannel)
+	}, rs.DeschedulingInterval, ctx.Done())
 
 	return nil
 }
+
+func getPluginConfig(pluginName string, pluginConfigs []api.PluginConfig) *api.PluginConfig {
+	for _, pluginConfig := range pluginConfigs {
+		if pluginConfig.Name == pluginName {
+			return &pluginConfig
+		}
+	}
+	return nil
+}
+
+func createClients(kubeconfig string) (clientset.Interface, clientset.Interface, error) {
+	kClient, err := client.CreateClient(kubeconfig, "descheduler")
+	if err != nil {
+		return nil, nil, err
+	}
+
+	eventClient, err := client.CreateClient(kubeconfig, "")
+	if err != nil {
+		return nil, nil, err
+	}
+
+	return kClient, eventClient, nil
+}

pkg/descheduler/descheduler_test.go

@@ -3,20 +3,27 @@ package descheduler
 import (
 	"context"
 	"fmt"
-	"strings"
 	"testing"
 	"time"
 
 	v1 "k8s.io/api/core/v1"
+	policy "k8s.io/api/policy/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/util/wait"
+	"k8s.io/apimachinery/pkg/runtime"
 	fakeclientset "k8s.io/client-go/kubernetes/fake"
+	core "k8s.io/client-go/testing"
 	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"sigs.k8s.io/descheduler/pkg/api"
+	"sigs.k8s.io/descheduler/pkg/api/v1alpha1"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/pluginbuilder"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removeduplicates"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingnodetaints"
 	"sigs.k8s.io/descheduler/test"
 )
 
 func TestTaintsUpdated(t *testing.T) {
+	pluginbuilder.PluginRegistry = pluginbuilder.NewRegistry()
+	pluginbuilder.Register(removepodsviolatingnodetaints.PluginName, removepodsviolatingnodetaints.New, &removepodsviolatingnodetaints.RemovePodsViolatingNodeTaintsArgs{}, pluginbuilder.PluginRegistry)
 	ctx := context.Background()
 	n1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
 	n2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
@@ -27,37 +34,21 @@ func TestTaintsUpdated(t *testing.T) {
 	}
 
 	client := fakeclientset.NewSimpleClientset(n1, n2, p1)
-	dp := &api.DeschedulerPolicy{
-		Strategies: api.StrategyList{
-			"RemovePodsViolatingNodeTaints": api.DeschedulerStrategy{
+	eventClient := fakeclientset.NewSimpleClientset(n1, n2, p1)
+	dp := &v1alpha1.DeschedulerPolicy{
+		Strategies: v1alpha1.StrategyList{
+			"RemovePodsViolatingNodeTaints": v1alpha1.DeschedulerStrategy{
 				Enabled: true,
 			},
 		},
 	}
 
-	stopChannel := make(chan struct{})
-	defer close(stopChannel)
-
 	rs, err := options.NewDeschedulerServer()
 	if err != nil {
 		t.Fatalf("Unable to initialize server: %v", err)
 	}
 	rs.Client = client
-	rs.DeschedulingInterval = 100 * time.Millisecond
-	errChan := make(chan error, 1)
-	defer close(errChan)
-	go func() {
-		err := RunDeschedulerStrategies(ctx, rs, dp, "v1beta1", stopChannel)
-		errChan <- err
-	}()
-	select {
-	case err := <-errChan:
-		if err != nil {
-			t.Fatalf("Unable to run descheduler strategies: %v", err)
-		}
-	case <-time.After(300 * time.Millisecond):
-		// Wait for few cycles and then verify the only pod still exists
-	}
+	rs.EventClient = eventClient
 
 	pods, err := client.CoreV1().Pods(p1.Namespace).List(ctx, metav1.ListOptions{})
 	if err != nil {
@@ -80,24 +71,165 @@ func TestTaintsUpdated(t *testing.T) {
 		t.Fatalf("Unable to update node: %v\n", err)
 	}
 
-	if err := wait.PollImmediate(100*time.Millisecond, time.Second, func() (bool, error) {
-		// Get over evicted pod result in panic
-		//pods, err := client.CoreV1().Pods(p1.Namespace).Get(p1.Name, metav1.GetOptions{})
-		// List is better, it does not panic.
-		// Though once the pod is evicted, List starts to error with "can't assign or convert v1beta1.Eviction into v1.Pod"
-		pods, err := client.CoreV1().Pods(p1.Namespace).List(ctx, metav1.ListOptions{})
-		if err == nil {
-			if len(pods.Items) > 0 {
-				return false, nil
-			}
-			return true, nil
-		}
-		if strings.Contains(err.Error(), "can't assign or convert v1beta1.Eviction into v1.Pod") {
-			return true, nil
-		}
+	var evictedPods []string
+	client.PrependReactor("create", "pods", podEvictionReactionFuc(&evictedPods))
 
-		return false, nil
-	}); err != nil {
-		t.Fatalf("Unable to evict pod, node taint did not get propagated to descheduler strategies")
+	internalDeschedulerPolicy, err := V1alpha1ToInternal(client, dp, pluginbuilder.PluginRegistry)
+	if err != nil {
+		t.Fatalf("Unable to convert v1alpha1 to v1alpha2: %v", err)
+	}
+
+	if err := RunDeschedulerStrategies(ctx, rs, internalDeschedulerPolicy, "v1"); err != nil {
+		t.Fatalf("Unable to run descheduler strategies: %v", err)
+	}
+
+	if len(evictedPods) != 1 {
+		t.Fatalf("Unable to evict pod, node taint did not get propagated to descheduler strategies %v\n", err)
+	}
+}
+
+func TestDuplicate(t *testing.T) {
+	pluginbuilder.PluginRegistry = pluginbuilder.NewRegistry()
+	pluginbuilder.Register(removeduplicates.PluginName, removeduplicates.New, &removeduplicates.RemoveDuplicatesArgs{}, pluginbuilder.PluginRegistry)
+	ctx := context.Background()
+	node1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
+	node2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
+
+	p1 := test.BuildTestPod("p1", 100, 0, node1.Name, nil)
+	p1.Namespace = "dev"
+	p2 := test.BuildTestPod("p2", 100, 0, node1.Name, nil)
+	p2.Namespace = "dev"
+	p3 := test.BuildTestPod("p3", 100, 0, node1.Name, nil)
+	p3.Namespace = "dev"
+
+	ownerRef1 := test.GetReplicaSetOwnerRefList()
+	p1.ObjectMeta.OwnerReferences = ownerRef1
+	p2.ObjectMeta.OwnerReferences = ownerRef1
+	p3.ObjectMeta.OwnerReferences = ownerRef1
+
+	client := fakeclientset.NewSimpleClientset(node1, node2, p1, p2, p3)
+	eventClient := fakeclientset.NewSimpleClientset(node1, node2, p1, p2, p3)
+	dp := &v1alpha1.DeschedulerPolicy{
+		Strategies: v1alpha1.StrategyList{
+			"RemoveDuplicates": v1alpha1.DeschedulerStrategy{
+				Enabled: true,
+			},
+		},
+	}
+
+	rs, err := options.NewDeschedulerServer()
+	if err != nil {
+		t.Fatalf("Unable to initialize server: %v", err)
+	}
+	rs.Client = client
+	rs.EventClient = eventClient
+
+	pods, err := client.CoreV1().Pods(p1.Namespace).List(ctx, metav1.ListOptions{})
+	if err != nil {
+		t.Errorf("Unable to list pods: %v", err)
+	}
+
+	if len(pods.Items) != 3 {
+		t.Errorf("Pods number should be 3 before evict")
+	}
+
+	var evictedPods []string
+	client.PrependReactor("create", "pods", podEvictionReactionFuc(&evictedPods))
+
+	internalDeschedulerPolicy, err := V1alpha1ToInternal(client, dp, pluginbuilder.PluginRegistry)
+	if err != nil {
+		t.Fatalf("Unable to convert v1alpha1 to v1alpha2: %v", err)
+	}
+	if err := RunDeschedulerStrategies(ctx, rs, internalDeschedulerPolicy, "v1"); err != nil {
+		t.Fatalf("Unable to run descheduler strategies: %v", err)
+	}
+
+	if len(evictedPods) == 0 {
+		t.Fatalf("Unable to evict pod, node taint did not get propagated to descheduler strategies %v\n", err)
+	}
+}
+
+func TestRootCancel(t *testing.T) {
+	ctx, cancel := context.WithCancel(context.Background())
+	n1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
+	n2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
+	client := fakeclientset.NewSimpleClientset(n1, n2)
+	eventClient := fakeclientset.NewSimpleClientset(n1, n2)
+	dp := &api.DeschedulerPolicy{
+		Profiles: []api.Profile{}, // no strategies needed for this test
+	}
+
+	rs, err := options.NewDeschedulerServer()
+	if err != nil {
+		t.Fatalf("Unable to initialize server: %v", err)
+	}
+	rs.Client = client
+	rs.EventClient = eventClient
+	rs.DeschedulingInterval = 100 * time.Millisecond
+	errChan := make(chan error, 1)
+	defer close(errChan)
+
+	go func() {
+		err := RunDeschedulerStrategies(ctx, rs, dp, "v1")
+		errChan <- err
+	}()
+	cancel()
+	select {
+	case err := <-errChan:
+		if err != nil {
+			t.Fatalf("Unable to run descheduler strategies: %v", err)
+		}
+	case <-time.After(1 * time.Second):
+		t.Fatal("Root ctx should have canceled immediately")
+	}
+}
+
+func TestRootCancelWithNoInterval(t *testing.T) {
+	ctx, cancel := context.WithCancel(context.Background())
+	n1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
+	n2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
+	client := fakeclientset.NewSimpleClientset(n1, n2)
+	eventClient := fakeclientset.NewSimpleClientset(n1, n2)
+	dp := &api.DeschedulerPolicy{
+		Profiles: []api.Profile{}, // no strategies needed for this test
+	}
+
+	rs, err := options.NewDeschedulerServer()
+	if err != nil {
+		t.Fatalf("Unable to initialize server: %v", err)
+	}
+	rs.Client = client
+	rs.EventClient = eventClient
+	rs.DeschedulingInterval = 0
+	errChan := make(chan error, 1)
+	defer close(errChan)
+
+	go func() {
+		err := RunDeschedulerStrategies(ctx, rs, dp, "v1")
+		errChan <- err
+	}()
+	cancel()
+	select {
+	case err := <-errChan:
+		if err != nil {
+			t.Fatalf("Unable to run descheduler strategies: %v", err)
+		}
+	case <-time.After(1 * time.Second):
+		t.Fatal("Root ctx should have canceled immediately")
+	}
+}
+
+func podEvictionReactionFuc(evictedPods *[]string) func(action core.Action) (bool, runtime.Object, error) {
+	return func(action core.Action) (bool, runtime.Object, error) {
+		if action.GetSubresource() == "eviction" {
+			createAct, matched := action.(core.CreateActionImpl)
+			if !matched {
+				return false, nil, fmt.Errorf("unable to convert action to core.CreateActionImpl")
+			}
+			if eviction, matched := createAct.Object.(*policy.Eviction); matched {
+				*evictedPods = append(*evictedPods, eviction.GetName())
+			}
+		}
+		return false, nil, nil // fallback to the default reactor
 	}
 }

pkg/descheduler/evictions/evictions.go

@@ -19,129 +19,160 @@ package evictions
 import (
 	"context"
 	"fmt"
-	"strings"
 
 	v1 "k8s.io/api/core/v1"
-	policy "k8s.io/api/policy/v1beta1"
+	policy "k8s.io/api/policy/v1"
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/labels"
-	"k8s.io/apimachinery/pkg/util/errors"
 	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/client-go/kubernetes/scheme"
-	clientcorev1 "k8s.io/client-go/kubernetes/typed/core/v1"
-	"k8s.io/client-go/tools/record"
+	"k8s.io/client-go/tools/events"
 	"k8s.io/klog/v2"
 	"sigs.k8s.io/descheduler/metrics"
-	nodeutil "sigs.k8s.io/descheduler/pkg/descheduler/node"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/utils"
 
 	eutils "sigs.k8s.io/descheduler/pkg/descheduler/evictions/utils"
 )
 
-const (
-	evictPodAnnotationKey = "descheduler.alpha.kubernetes.io/evict"
+// nodePodEvictedCount keeps count of pods evicted on node
+type (
+	nodePodEvictedCount    map[string]uint
+	namespacePodEvictCount map[string]uint
 )
 
-// nodePodEvictedCount keeps count of pods evicted on node
-type nodePodEvictedCount map[*v1.Node]int
-
 type PodEvictor struct {
-	client                  clientset.Interface
-	nodes                   []*v1.Node
-	policyGroupVersion      string
-	dryRun                  bool
-	maxPodsToEvictPerNode   int
-	nodepodCount            nodePodEvictedCount
-	evictLocalStoragePods   bool
-	evictSystemCriticalPods bool
-	ignorePvcPods           bool
+	client                     clientset.Interface
+	nodes                      []*v1.Node
+	policyGroupVersion         string
+	dryRun                     bool
+	maxPodsToEvictPerNode      *uint
+	maxPodsToEvictPerNamespace *uint
+	nodepodCount               nodePodEvictedCount
+	namespacePodCount          namespacePodEvictCount
+	metricsEnabled             bool
+	eventRecorder              events.EventRecorder
 }
 
 func NewPodEvictor(
 	client clientset.Interface,
 	policyGroupVersion string,
 	dryRun bool,
-	maxPodsToEvictPerNode int,
+	maxPodsToEvictPerNode *uint,
+	maxPodsToEvictPerNamespace *uint,
 	nodes []*v1.Node,
-	evictLocalStoragePods bool,
-	evictSystemCriticalPods bool,
-	ignorePvcPods bool,
+	metricsEnabled bool,
+	eventRecorder events.EventRecorder,
 ) *PodEvictor {
-	var nodePodCount = make(nodePodEvictedCount)
+	nodePodCount := make(nodePodEvictedCount)
+	namespacePodCount := make(namespacePodEvictCount)
 	for _, node := range nodes {
 		// Initialize podsEvicted till now with 0.
-		nodePodCount[node] = 0
+		nodePodCount[node.Name] = 0
 	}
 
 	return &PodEvictor{
-		client:                  client,
-		nodes:                   nodes,
-		policyGroupVersion:      policyGroupVersion,
-		dryRun:                  dryRun,
-		maxPodsToEvictPerNode:   maxPodsToEvictPerNode,
-		nodepodCount:            nodePodCount,
-		evictLocalStoragePods:   evictLocalStoragePods,
-		evictSystemCriticalPods: evictSystemCriticalPods,
-		ignorePvcPods:           ignorePvcPods,
+		client:                     client,
+		nodes:                      nodes,
+		policyGroupVersion:         policyGroupVersion,
+		dryRun:                     dryRun,
+		maxPodsToEvictPerNode:      maxPodsToEvictPerNode,
+		maxPodsToEvictPerNamespace: maxPodsToEvictPerNamespace,
+		nodepodCount:               nodePodCount,
+		namespacePodCount:          namespacePodCount,
+		metricsEnabled:             metricsEnabled,
+		eventRecorder:              eventRecorder,
 	}
 }
 
 // NodeEvicted gives a number of pods evicted for node
-func (pe *PodEvictor) NodeEvicted(node *v1.Node) int {
-	return pe.nodepodCount[node]
+func (pe *PodEvictor) NodeEvicted(node *v1.Node) uint {
+	return pe.nodepodCount[node.Name]
 }
 
 // TotalEvicted gives a number of pods evicted through all nodes
-func (pe *PodEvictor) TotalEvicted() int {
-	var total int
+func (pe *PodEvictor) TotalEvicted() uint {
+	var total uint
 	for _, count := range pe.nodepodCount {
 		total += count
 	}
 	return total
 }
 
-// EvictPod returns non-nil error only when evicting a pod on a node is not
-// possible (due to maxPodsToEvictPerNode constraint). Success is true when the pod
-// is evicted on the server side.
-func (pe *PodEvictor) EvictPod(ctx context.Context, pod *v1.Pod, node *v1.Node, strategy string, reasons ...string) (bool, error) {
-	reason := strategy
-	if len(reasons) > 0 {
-		reason += " (" + strings.Join(reasons, ", ") + ")"
+// NodeLimitExceeded checks if the number of evictions for a node was exceeded
+func (pe *PodEvictor) NodeLimitExceeded(node *v1.Node) bool {
+	if pe.maxPodsToEvictPerNode != nil {
+		return pe.nodepodCount[node.Name] == *pe.maxPodsToEvictPerNode
 	}
-	if pe.maxPodsToEvictPerNode > 0 && pe.nodepodCount[node]+1 > pe.maxPodsToEvictPerNode {
-		metrics.PodsEvicted.With(map[string]string{"result": "maximum number reached", "strategy": strategy, "namespace": pod.Namespace}).Inc()
-		return false, fmt.Errorf("Maximum number %v of evicted pods per %q node reached", pe.maxPodsToEvictPerNode, node.Name)
-	}
-
-	err := evictPod(ctx, pe.client, pod, pe.policyGroupVersion, pe.dryRun)
-	if err != nil {
-		// err is used only for logging purposes
-		klog.ErrorS(err, "Error evicting pod", "pod", klog.KObj(pod), "reason", reason)
-		metrics.PodsEvicted.With(map[string]string{"result": "error", "strategy": strategy, "namespace": pod.Namespace}).Inc()
-		return false, nil
-	}
-
-	pe.nodepodCount[node]++
-	if pe.dryRun {
-		klog.V(1).InfoS("Evicted pod in dry run mode", "pod", klog.KObj(pod), "reason", reason)
-	} else {
-		klog.V(1).InfoS("Evicted pod", "pod", klog.KObj(pod), "reason", reason)
-		eventBroadcaster := record.NewBroadcaster()
-		eventBroadcaster.StartStructuredLogging(3)
-		eventBroadcaster.StartRecordingToSink(&clientcorev1.EventSinkImpl{Interface: pe.client.CoreV1().Events(pod.Namespace)})
-		r := eventBroadcaster.NewRecorder(scheme.Scheme, v1.EventSource{Component: "sigs.k8s.io.descheduler"})
-		r.Event(pod, v1.EventTypeNormal, "Descheduled", fmt.Sprintf("pod evicted by sigs.k8s.io/descheduler%s", reason))
-		metrics.PodsEvicted.With(map[string]string{"result": "success", "strategy": strategy, "namespace": pod.Namespace}).Inc()
-	}
-	return true, nil
+	return false
 }
 
-func evictPod(ctx context.Context, client clientset.Interface, pod *v1.Pod, policyGroupVersion string, dryRun bool) error {
-	if dryRun {
-		return nil
+// EvictOptions provides a handle for passing additional info to EvictPod
+type EvictOptions struct {
+	// Reason allows for passing details about the specific eviction for logging.
+	Reason string
+}
+
+// EvictPod evicts a pod while exercising eviction limits.
+// Returns true when the pod is evicted on the server side.
+func (pe *PodEvictor) EvictPod(ctx context.Context, pod *v1.Pod, opts EvictOptions) bool {
+	// TODO: Replace context-propagated Strategy name with a defined framework handle for accessing Strategy info
+	strategy := ""
+	if ctx.Value("strategyName") != nil {
+		strategy = ctx.Value("strategyName").(string)
 	}
+
+	if pod.Spec.NodeName != "" {
+		if pe.maxPodsToEvictPerNode != nil && pe.nodepodCount[pod.Spec.NodeName]+1 > *pe.maxPodsToEvictPerNode {
+			if pe.metricsEnabled {
+				metrics.PodsEvicted.With(map[string]string{"result": "maximum number of pods per node reached", "strategy": strategy, "namespace": pod.Namespace, "node": pod.Spec.NodeName}).Inc()
+			}
+			klog.ErrorS(fmt.Errorf("Maximum number of evicted pods per node reached"), "limit", *pe.maxPodsToEvictPerNode, "node", pod.Spec.NodeName)
+			return false
+		}
+	}
+
+	if pe.maxPodsToEvictPerNamespace != nil && pe.namespacePodCount[pod.Namespace]+1 > *pe.maxPodsToEvictPerNamespace {
+		if pe.metricsEnabled {
+			metrics.PodsEvicted.With(map[string]string{"result": "maximum number of pods per namespace reached", "strategy": strategy, "namespace": pod.Namespace, "node": pod.Spec.NodeName}).Inc()
+		}
+		klog.ErrorS(fmt.Errorf("Maximum number of evicted pods per namespace reached"), "limit", *pe.maxPodsToEvictPerNamespace, "namespace", pod.Namespace)
+		return false
+	}
+
+	err := evictPod(ctx, pe.client, pod, pe.policyGroupVersion)
+	if err != nil {
+		// err is used only for logging purposes
+		klog.ErrorS(err, "Error evicting pod", "pod", klog.KObj(pod), "reason", opts.Reason)
+		if pe.metricsEnabled {
+			metrics.PodsEvicted.With(map[string]string{"result": "error", "strategy": strategy, "namespace": pod.Namespace, "node": pod.Spec.NodeName}).Inc()
+		}
+		return false
+	}
+
+	if pod.Spec.NodeName != "" {
+		pe.nodepodCount[pod.Spec.NodeName]++
+	}
+	pe.namespacePodCount[pod.Namespace]++
+
+	if pe.metricsEnabled {
+		metrics.PodsEvicted.With(map[string]string{"result": "success", "strategy": strategy, "namespace": pod.Namespace, "node": pod.Spec.NodeName}).Inc()
+	}
+
+	if pe.dryRun {
+		klog.V(1).InfoS("Evicted pod in dry run mode", "pod", klog.KObj(pod), "reason", opts.Reason, "strategy", strategy, "node", pod.Spec.NodeName)
+	} else {
+		klog.V(1).InfoS("Evicted pod", "pod", klog.KObj(pod), "reason", opts.Reason, "strategy", strategy, "node", pod.Spec.NodeName)
+		reason := opts.Reason
+		if len(reason) == 0 {
+			reason = strategy
+			if len(reason) == 0 {
+				reason = "NotSet"
+			}
+		}
+		pe.eventRecorder.Eventf(pod, nil, v1.EventTypeNormal, reason, "Descheduled", "pod evicted from %v node by sigs.k8s.io/descheduler", pod.Spec.NodeName)
+	}
+	return true
+}
+
+func evictPod(ctx context.Context, client clientset.Interface, pod *v1.Pod, policyGroupVersion string) error {
 	deleteOptions := &metav1.DeleteOptions{}
 	// GracePeriodSeconds ?
 	eviction := &policy.Eviction{
@@ -155,7 +186,7 @@ func evictPod(ctx context.Context, client clientset.Interface, pod *v1.Pod, poli
 		},
 		DeleteOptions: deleteOptions,
 	}
-	err := client.PolicyV1beta1().Evictions(eviction.Namespace).Evict(ctx, eviction)
+	err := client.PolicyV1().Evictions(eviction.Namespace).Evict(ctx, eviction)
 
 	if apierrors.IsTooManyRequests(err) {
 		return fmt.Errorf("error when evicting pod (ignoring) %q: %v", pod.Name, err)
@@ -165,152 +196,3 @@ func evictPod(ctx context.Context, client clientset.Interface, pod *v1.Pod, poli
 	}
 	return err
 }
-
-type Options struct {
-	priority      *int32
-	nodeFit       bool
-	labelSelector labels.Selector
-}
-
-// WithPriorityThreshold sets a threshold for pod's priority class.
-// Any pod whose priority class is lower is evictable.
-func WithPriorityThreshold(priority int32) func(opts *Options) {
-	return func(opts *Options) {
-		var p int32 = priority
-		opts.priority = &p
-	}
-}
-
-// WithNodeFit sets whether or not to consider taints, node selectors,
-// and pod affinity when evicting. A pod whose tolerations, node selectors,
-// and affinity match a node other than the one it is currently running on
-// is evictable.
-func WithNodeFit(nodeFit bool) func(opts *Options) {
-	return func(opts *Options) {
-		opts.nodeFit = nodeFit
-	}
-}
-
-// WithLabelSelector sets whether or not to apply label filtering when evicting.
-// Any pod matching the label selector is considered evictable.
-func WithLabelSelector(labelSelector labels.Selector) func(opts *Options) {
-	return func(opts *Options) {
-		opts.labelSelector = labelSelector
-	}
-}
-
-type constraint func(pod *v1.Pod) error
-
-type evictable struct {
-	constraints []constraint
-}
-
-// Evictable provides an implementation of IsEvictable(IsEvictable(pod *v1.Pod) bool).
-// The method accepts a list of options which allow to extend constraints
-// which decides when a pod is considered evictable.
-func (pe *PodEvictor) Evictable(opts ...func(opts *Options)) *evictable {
-	options := &Options{}
-	for _, opt := range opts {
-		opt(options)
-	}
-
-	ev := &evictable{}
-	if !pe.evictSystemCriticalPods {
-		ev.constraints = append(ev.constraints, func(pod *v1.Pod) error {
-			// Moved from IsEvictable function to allow for disabling
-			if utils.IsCriticalPriorityPod(pod) {
-				return fmt.Errorf("pod has system critical priority")
-			}
-			return nil
-		})
-
-		if options.priority != nil {
-			ev.constraints = append(ev.constraints, func(pod *v1.Pod) error {
-				if IsPodEvictableBasedOnPriority(pod, *options.priority) {
-					return nil
-				}
-				return fmt.Errorf("pod has higher priority than specified priority class threshold")
-			})
-		}
-	}
-	if !pe.evictLocalStoragePods {
-		ev.constraints = append(ev.constraints, func(pod *v1.Pod) error {
-			if utils.IsPodWithLocalStorage(pod) {
-				return fmt.Errorf("pod has local storage and descheduler is not configured with evictLocalStoragePods")
-			}
-			return nil
-		})
-	}
-	if pe.ignorePvcPods {
-		ev.constraints = append(ev.constraints, func(pod *v1.Pod) error {
-			if utils.IsPodWithPVC(pod) {
-				return fmt.Errorf("pod has a PVC and descheduler is configured to ignore PVC pods")
-			}
-			return nil
-		})
-	}
-	if options.nodeFit {
-		ev.constraints = append(ev.constraints, func(pod *v1.Pod) error {
-			if !nodeutil.PodFitsAnyOtherNode(pod, pe.nodes) {
-				return fmt.Errorf("pod does not fit on any other node because of nodeSelector(s), Taint(s), or nodes marked as unschedulable")
-			}
-			return nil
-		})
-	}
-	if options.labelSelector != nil && !options.labelSelector.Empty() {
-		ev.constraints = append(ev.constraints, func(pod *v1.Pod) error {
-			if !options.labelSelector.Matches(labels.Set(pod.Labels)) {
-				return fmt.Errorf("pod labels do not match the labelSelector filter in the policy parameter")
-			}
-			return nil
-		})
-	}
-
-	return ev
-}
-
-// IsEvictable decides when a pod is evictable
-func (ev *evictable) IsEvictable(pod *v1.Pod) bool {
-	checkErrs := []error{}
-
-	ownerRefList := podutil.OwnerRef(pod)
-	if utils.IsDaemonsetPod(ownerRefList) {
-		checkErrs = append(checkErrs, fmt.Errorf("pod is a DaemonSet pod"))
-	}
-
-	if len(ownerRefList) == 0 {
-		checkErrs = append(checkErrs, fmt.Errorf("pod does not have any ownerrefs"))
-	}
-
-	if utils.IsMirrorPod(pod) {
-		checkErrs = append(checkErrs, fmt.Errorf("pod is a mirror pod"))
-	}
-
-	if utils.IsStaticPod(pod) {
-		checkErrs = append(checkErrs, fmt.Errorf("pod is a static pod"))
-	}
-
-	for _, c := range ev.constraints {
-		if err := c(pod); err != nil {
-			checkErrs = append(checkErrs, err)
-		}
-	}
-
-	if len(checkErrs) > 0 && !HaveEvictAnnotation(pod) {
-		klog.V(4).InfoS("Pod lacks an eviction annotation and fails the following checks", "pod", klog.KObj(pod), "checks", errors.NewAggregate(checkErrs).Error())
-		return false
-	}
-
-	return true
-}
-
-// HaveEvictAnnotation checks if the pod have evict annotation
-func HaveEvictAnnotation(pod *v1.Pod) bool {
-	_, found := pod.ObjectMeta.Annotations[evictPodAnnotationKey]
-	return found
-}
-
-// IsPodEvictableBasedOnPriority checks if the given pod is evictable based on priority resolved from pod Spec.
-func IsPodEvictableBasedOnPriority(pod *v1.Pod, priority int32) bool {
-	return pod.Spec.Priority == nil || *pod.Spec.Priority < priority
-}

pkg/descheduler/evictions/evictions_test.go

@@ -62,381 +62,13 @@ func TestEvictPod(t *testing.T) {
 		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
 			return true, &v1.PodList{Items: test.pods}, nil
 		})
-		got := evictPod(ctx, fakeClient, test.pod, "v1", false)
+		got := evictPod(ctx, fakeClient, test.pod, "v1")
 		if got != test.want {
 			t.Errorf("Test error for Desc: %s. Expected %v pod eviction to be %v, got %v", test.description, test.pod.Name, test.want, got)
 		}
 	}
 }
 
-func TestIsEvictable(t *testing.T) {
-	n1 := test.BuildTestNode("node1", 1000, 2000, 13, nil)
-	lowPriority := int32(800)
-	highPriority := int32(900)
-
-	nodeTaintKey := "hardware"
-	nodeTaintValue := "gpu"
-
-	nodeLabelKey := "datacenter"
-	nodeLabelValue := "east"
-	type testCase struct {
-		pod                     *v1.Pod
-		nodes                   []*v1.Node
-		runBefore               func(*v1.Pod, []*v1.Node)
-		evictLocalStoragePods   bool
-		evictSystemCriticalPods bool
-		priorityThreshold       *int32
-		nodeFit                 bool
-		result                  bool
-	}
-
-	testCases := []testCase{
-		{ // Normal pod eviction with normal ownerRefs
-			pod: test.BuildTestPod("p1", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Normal pod eviction with normal ownerRefs and descheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p2", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Normal pod eviction with replicaSet ownerRefs
-			pod: test.BuildTestPod("p3", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Normal pod eviction with replicaSet ownerRefs and descheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p4", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Normal pod eviction with statefulSet ownerRefs
-			pod: test.BuildTestPod("p18", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetStatefulSetOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Normal pod eviction with statefulSet ownerRefs and descheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p19", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.ObjectMeta.OwnerReferences = test.GetStatefulSetOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Pod not evicted because it is bound to a PV and evictLocalStoragePods = false
-			pod: test.BuildTestPod("p5", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Spec.Volumes = []v1.Volume{
-					{
-						Name: "sample",
-						VolumeSource: v1.VolumeSource{
-							HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-							EmptyDir: &v1.EmptyDirVolumeSource{
-								SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
-						},
-					},
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  false,
-		}, { // Pod is evicted because it is bound to a PV and evictLocalStoragePods = true
-			pod: test.BuildTestPod("p6", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Spec.Volumes = []v1.Volume{
-					{
-						Name: "sample",
-						VolumeSource: v1.VolumeSource{
-							HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-							EmptyDir: &v1.EmptyDirVolumeSource{
-								SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
-						},
-					},
-				}
-			},
-			evictLocalStoragePods:   true,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Pod is evicted because it is bound to a PV and evictLocalStoragePods = false, but it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p7", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Spec.Volumes = []v1.Volume{
-					{
-						Name: "sample",
-						VolumeSource: v1.VolumeSource{
-							HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-							EmptyDir: &v1.EmptyDirVolumeSource{
-								SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
-						},
-					},
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Pod not evicted becasuse it is part of a daemonSet
-			pod: test.BuildTestPod("p8", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  false,
-		}, { // Pod is evicted becasuse it is part of a daemonSet, but it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p9", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Pod not evicted becasuse it is a mirror pod
-			pod: test.BuildTestPod("p10", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Annotations = test.GetMirrorPodAnnotation()
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  false,
-		}, { // Pod is evicted becasuse it is a mirror pod, but it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p11", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Annotations = test.GetMirrorPodAnnotation()
-				pod.Annotations["descheduler.alpha.kubernetes.io/evict"] = "true"
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Pod not evicted becasuse it has system critical priority
-			pod: test.BuildTestPod("p12", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				priority := utils.SystemCriticalPriority
-				pod.Spec.Priority = &priority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  false,
-		}, { // Pod is evicted becasuse it has system critical priority, but it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p13", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				priority := utils.SystemCriticalPriority
-				pod.Spec.Priority = &priority
-				pod.Annotations = map[string]string{
-					"descheduler.alpha.kubernetes.io/evict": "true",
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			result:                  true,
-		}, { // Pod not evicted becasuse it has a priority higher than the configured priority threshold
-			pod: test.BuildTestPod("p14", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Spec.Priority = &highPriority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			priorityThreshold:       &lowPriority,
-			result:                  false,
-		}, { // Pod is evicted becasuse it has a priority higher than the configured priority threshold, but it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p15", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.Spec.Priority = &highPriority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			priorityThreshold:       &lowPriority,
-			result:                  true,
-		}, { // Pod is evicted becasuse it has system critical priority, but evictSystemCriticalPods = true
-			pod: test.BuildTestPod("p16", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				priority := utils.SystemCriticalPriority
-				pod.Spec.Priority = &priority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: true,
-			result:                  true,
-		}, { // Pod is evicted becasuse it has system critical priority, but evictSystemCriticalPods = true and it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p16", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				priority := utils.SystemCriticalPriority
-				pod.Spec.Priority = &priority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: true,
-			result:                  true,
-		}, { // Pod is evicted becasuse it has a priority higher than the configured priority threshold, but evictSystemCriticalPods = true
-			pod: test.BuildTestPod("p17", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Spec.Priority = &highPriority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: true,
-			priorityThreshold:       &lowPriority,
-			result:                  true,
-		}, { // Pod is evicted becasuse it has a priority higher than the configured priority threshold, but evictSystemCriticalPods = true and it has scheduler.alpha.kubernetes.io/evict annotation
-			pod: test.BuildTestPod("p17", 400, 0, n1.Name, nil),
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
-				pod.Spec.Priority = &highPriority
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: true,
-			priorityThreshold:       &lowPriority,
-			result:                  true,
-		}, { // Pod with no tolerations running on normal node, all other nodes tainted
-			pod:   test.BuildTestPod("p1", 400, 0, n1.Name, nil),
-			nodes: []*v1.Node{test.BuildTestNode("node2", 1000, 2000, 13, nil), test.BuildTestNode("node3", 1000, 2000, 13, nil)},
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-				for _, node := range nodes {
-					node.Spec.Taints = []v1.Taint{
-						{
-							Key:    nodeTaintKey,
-							Value:  nodeTaintValue,
-							Effect: v1.TaintEffectNoSchedule,
-						},
-					}
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			nodeFit:                 true,
-			result:                  false,
-		}, { // Pod with correct tolerations running on normal node, all other nodes tainted
-			pod: test.BuildTestPod("p1", 400, 0, n1.Name, func(pod *v1.Pod) {
-				pod.Spec.Tolerations = []v1.Toleration{
-					{
-						Key:    nodeTaintKey,
-						Value:  nodeTaintValue,
-						Effect: v1.TaintEffectNoSchedule,
-					},
-				}
-			}),
-			nodes: []*v1.Node{test.BuildTestNode("node2", 1000, 2000, 13, nil), test.BuildTestNode("node3", 1000, 2000, 13, nil)},
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-				for _, node := range nodes {
-					node.Spec.Taints = []v1.Taint{
-						{
-							Key:    nodeTaintKey,
-							Value:  nodeTaintValue,
-							Effect: v1.TaintEffectNoSchedule,
-						},
-					}
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			nodeFit:                 true,
-			result:                  true,
-		}, { // Pod with incorrect node selector
-			pod: test.BuildTestPod("p1", 400, 0, n1.Name, func(pod *v1.Pod) {
-				pod.Spec.NodeSelector = map[string]string{
-					nodeLabelKey: "fail",
-				}
-			}),
-			nodes: []*v1.Node{test.BuildTestNode("node2", 1000, 2000, 13, nil), test.BuildTestNode("node3", 1000, 2000, 13, nil)},
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-				for _, node := range nodes {
-					node.ObjectMeta.Labels = map[string]string{
-						nodeLabelKey: nodeLabelValue,
-					}
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			nodeFit:                 true,
-			result:                  false,
-		}, { // Pod with correct node selector
-			pod: test.BuildTestPod("p1", 400, 0, n1.Name, func(pod *v1.Pod) {
-				pod.Spec.NodeSelector = map[string]string{
-					nodeLabelKey: nodeLabelValue,
-				}
-			}),
-			nodes: []*v1.Node{test.BuildTestNode("node2", 1000, 2000, 13, nil), test.BuildTestNode("node3", 1000, 2000, 13, nil)},
-			runBefore: func(pod *v1.Pod, nodes []*v1.Node) {
-				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-				for _, node := range nodes {
-					node.ObjectMeta.Labels = map[string]string{
-						nodeLabelKey: nodeLabelValue,
-					}
-				}
-			},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			nodeFit:                 true,
-			result:                  true,
-		},
-	}
-
-	for _, test := range testCases {
-		test.runBefore(test.pod, test.nodes)
-		nodes := append(test.nodes, n1)
-
-		podEvictor := &PodEvictor{
-			evictLocalStoragePods:   test.evictLocalStoragePods,
-			evictSystemCriticalPods: test.evictSystemCriticalPods,
-			nodes:                   nodes,
-		}
-
-		evictable := podEvictor.Evictable()
-		var opts []func(opts *Options)
-		if test.priorityThreshold != nil {
-			opts = append(opts, WithPriorityThreshold(*test.priorityThreshold))
-		}
-		if test.nodeFit {
-			opts = append(opts, WithNodeFit(true))
-		}
-		evictable = podEvictor.Evictable(opts...)
-
-		result := evictable.IsEvictable(test.pod)
-		if result != test.result {
-			t.Errorf("IsEvictable should return for pod %s %t, but it returns %t", test.pod.Name, test.result, result)
-		}
-
-	}
-}
 func TestPodTypes(t *testing.T) {
 	n1 := test.BuildTestNode("node1", 1000, 2000, 9, nil)
 	p1 := test.BuildTestPod("p1", 400, 0, n1.Name, nil)
@@ -449,7 +81,7 @@ func TestPodTypes(t *testing.T) {
 	p1.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
 	// The following 4 pods won't get evicted.
 	// A daemonset.
-	//p2.Annotations = test.GetDaemonSetAnnotation()
+	// p2.Annotations = test.GetDaemonSetAnnotation()
 	p2.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
 	// A pod with local storage.
 	p3.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
@@ -459,7 +91,8 @@ func TestPodTypes(t *testing.T) {
 			VolumeSource: v1.VolumeSource{
 				HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
 				EmptyDir: &v1.EmptyDirVolumeSource{
-					SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
+					SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI),
+				},
 			},
 		},
 	}
@@ -479,5 +112,4 @@ func TestPodTypes(t *testing.T) {
 	if utils.IsDaemonsetPod(ownerRefList) || utils.IsPodWithLocalStorage(p1) || utils.IsCriticalPriorityPod(p1) || utils.IsMirrorPod(p1) || utils.IsStaticPod(p1) {
 		t.Errorf("Expected p1 to be a normal pod.")
 	}
-
 }

pkg/descheduler/leaderelection.go

@@ -0,0 +1,100 @@
+/*
+Copyright 2022 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+   http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package descheduler
+
+import (
+	"context"
+	"fmt"
+	"os"
+
+	"k8s.io/apimachinery/pkg/util/uuid"
+	clientset "k8s.io/client-go/kubernetes"
+	"k8s.io/client-go/tools/leaderelection"
+	"k8s.io/client-go/tools/leaderelection/resourcelock"
+	componentbaseconfig "k8s.io/component-base/config"
+	"k8s.io/klog/v2"
+)
+
+// NewLeaderElection starts the leader election code loop
+func NewLeaderElection(
+	run func() error,
+	client clientset.Interface,
+	LeaderElectionConfig *componentbaseconfig.LeaderElectionConfiguration,
+	ctx context.Context,
+) error {
+	var id string
+
+	if hostname, err := os.Hostname(); err != nil {
+		// on errors, make sure we're unique
+		id = string(uuid.NewUUID())
+	} else {
+		// add a uniquifier so that two processes on the same host don't accidentally both become active
+		id = hostname + "_" + string(uuid.NewUUID())
+	}
+
+	klog.V(3).Infof("Assigned unique lease holder id: %s", id)
+
+	if len(LeaderElectionConfig.ResourceNamespace) == 0 {
+		return fmt.Errorf("namespace may not be empty")
+	}
+
+	if len(LeaderElectionConfig.ResourceName) == 0 {
+		return fmt.Errorf("name may not be empty")
+	}
+
+	lock, err := resourcelock.New(
+		LeaderElectionConfig.ResourceLock,
+		LeaderElectionConfig.ResourceNamespace,
+		LeaderElectionConfig.ResourceName,
+		client.CoreV1(),
+		client.CoordinationV1(),
+		resourcelock.ResourceLockConfig{
+			Identity: id,
+		},
+	)
+	if err != nil {
+		return fmt.Errorf("unable to create leader election lock: %v", err)
+	}
+
+	leaderelection.RunOrDie(ctx, leaderelection.LeaderElectionConfig{
+		Lock:            lock,
+		ReleaseOnCancel: true,
+		LeaseDuration:   LeaderElectionConfig.LeaseDuration.Duration,
+		RenewDeadline:   LeaderElectionConfig.RenewDeadline.Duration,
+		RetryPeriod:     LeaderElectionConfig.RetryPeriod.Duration,
+		Callbacks: leaderelection.LeaderCallbacks{
+			OnStartedLeading: func(ctx context.Context) {
+				klog.V(1).InfoS("Started leading")
+				err := run()
+				if err != nil {
+					klog.Error(err)
+				}
+			},
+			OnStoppedLeading: func() {
+				klog.V(1).InfoS("Leader lost")
+			},
+			OnNewLeader: func(identity string) {
+				// Just got the lock
+				if identity == id {
+					return
+				}
+				klog.V(1).Infof("New leader elected: %v", identity)
+			},
+		},
+	})
+	return nil
+}

pkg/descheduler/node/node.go

@@ -18,19 +18,22 @@ package node
 
 import (
 	"context"
+	"fmt"
 
 	v1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/labels"
-	coreinformers "k8s.io/client-go/informers/core/v1"
 	clientset "k8s.io/client-go/kubernetes"
+	listersv1 "k8s.io/client-go/listers/core/v1"
 	"k8s.io/klog/v2"
+	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
 	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
 // ReadyNodes returns ready nodes irrespective of whether they are
 // schedulable or not.
-func ReadyNodes(ctx context.Context, client clientset.Interface, nodeInformer coreinformers.NodeInformer, nodeSelector string) ([]*v1.Node, error) {
+func ReadyNodes(ctx context.Context, client clientset.Interface, nodeLister listersv1.NodeLister, nodeSelector string) ([]*v1.Node, error) {
 	ns, err := labels.Parse(nodeSelector)
 	if err != nil {
 		return []*v1.Node{}, err
@@ -38,7 +41,7 @@ func ReadyNodes(ctx context.Context, client clientset.Interface, nodeInformer co
 
 	var nodes []*v1.Node
 	// err is defined above
-	if nodes, err = nodeInformer.Lister().List(ns); err != nil {
+	if nodes, err = nodeLister.List(ns); err != nil {
 		return []*v1.Node{}, err
 	}
 
@@ -96,32 +99,92 @@ func IsReady(node *v1.Node) bool {
 	return true
 }
 
-// PodFitsAnyOtherNode checks if the given pod fits any of the given nodes, besides the node
-// the pod is already running on. The node fit is based on multiple criteria, like, pod node selector
-// matching the node label (including affinity), the taints on the node, and the node being schedulable or not.
-func PodFitsAnyOtherNode(pod *v1.Pod, nodes []*v1.Node) bool {
+// NodeFit returns true if the provided pod can be scheduled onto the provided node.
+// This function is used when the NodeFit pod filtering feature of the Descheduler is enabled.
+// This function currently considers a subset of the Kubernetes Scheduler's predicates when
+// deciding if a pod would fit on a node, but more predicates may be added in the future.
+func NodeFit(nodeIndexer podutil.GetPodsAssignedToNodeFunc, pod *v1.Pod, node *v1.Node) []error {
+	// Check node selector and required affinity
+	var errors []error
+	if ok, err := utils.PodMatchNodeSelector(pod, node); err != nil {
+		errors = append(errors, err)
+	} else if !ok {
+		errors = append(errors, fmt.Errorf("pod node selector does not match the node label"))
+	}
+	// Check taints (we only care about NoSchedule and NoExecute taints)
+	ok := utils.TolerationsTolerateTaintsWithFilter(pod.Spec.Tolerations, node.Spec.Taints, func(taint *v1.Taint) bool {
+		return taint.Effect == v1.TaintEffectNoSchedule || taint.Effect == v1.TaintEffectNoExecute
+	})
+	if !ok {
+		errors = append(errors, fmt.Errorf("pod does not tolerate taints on the node"))
+	}
+	// Check if the pod can fit on a node based off it's requests
+	if pod.Spec.NodeName == "" || pod.Spec.NodeName != node.Name {
+		ok, reqErrors := fitsRequest(nodeIndexer, pod, node)
+		if !ok {
+			errors = append(errors, reqErrors...)
+		}
+	}
+	// Check if node is schedulable
+	if IsNodeUnschedulable(node) {
+		errors = append(errors, fmt.Errorf("node is not schedulable"))
+	}
 
+	return errors
+}
+
+// PodFitsAnyOtherNode checks if the given pod will fit any of the given nodes, besides the node
+// the pod is already running on. The predicates used to determine if the pod will fit can be found in the NodeFit function.
+func PodFitsAnyOtherNode(nodeIndexer podutil.GetPodsAssignedToNodeFunc, pod *v1.Pod, nodes []*v1.Node) bool {
 	for _, node := range nodes {
 		// Skip node pod is already on
 		if node.Name == pod.Spec.NodeName {
 			continue
 		}
-		// Check node selector and required affinity
-		ok, err := utils.PodMatchNodeSelector(pod, node)
-		if err != nil || !ok {
-			continue
-		}
-		// Check taints (we only care about NoSchedule and NoExecute taints)
-		ok = utils.TolerationsTolerateTaintsWithFilter(pod.Spec.Tolerations, node.Spec.Taints, func(taint *v1.Taint) bool {
-			return taint.Effect == v1.TaintEffectNoSchedule || taint.Effect == v1.TaintEffectNoExecute
-		})
-		if !ok {
-			continue
-		}
-		// Check if node is schedulable
-		if !IsNodeUnschedulable(node) {
-			klog.V(2).InfoS("Pod can possibly be scheduled on a different node", "pod", klog.KObj(pod), "node", klog.KObj(node))
+
+		errors := NodeFit(nodeIndexer, pod, node)
+		if len(errors) == 0 {
+			klog.V(4).InfoS("Pod fits on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
 			return true
+		} else {
+			klog.V(4).InfoS("Pod does not fit on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
+			for _, err := range errors {
+				klog.V(4).InfoS(err.Error())
+			}
+		}
+	}
+	return false
+}
+
+// PodFitsAnyNode checks if the given pod will fit any of the given nodes. The predicates used
+// to determine if the pod will fit can be found in the NodeFit function.
+func PodFitsAnyNode(nodeIndexer podutil.GetPodsAssignedToNodeFunc, pod *v1.Pod, nodes []*v1.Node) bool {
+	for _, node := range nodes {
+		errors := NodeFit(nodeIndexer, pod, node)
+		if len(errors) == 0 {
+			klog.V(4).InfoS("Pod fits on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
+			return true
+		} else {
+			klog.V(4).InfoS("Pod does not fit on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
+			for _, err := range errors {
+				klog.V(4).InfoS(err.Error())
+			}
+		}
+	}
+	return false
+}
+
+// PodFitsCurrentNode checks if the given pod will fit onto the given node. The predicates used
+// to determine if the pod will fit can be found in the NodeFit function.
+func PodFitsCurrentNode(nodeIndexer podutil.GetPodsAssignedToNodeFunc, pod *v1.Pod, node *v1.Node) bool {
+	errors := NodeFit(nodeIndexer, pod, node)
+	if len(errors) == 0 {
+		klog.V(4).InfoS("Pod fits on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
+		return true
+	} else {
+		klog.V(4).InfoS("Pod does not fit on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
+		for _, err := range errors {
+			klog.V(4).InfoS(err.Error())
 		}
 	}
 	return false
@@ -133,39 +196,95 @@ func IsNodeUnschedulable(node *v1.Node) bool {
 	return node.Spec.Unschedulable
 }
 
-// PodFitsAnyNode checks if the given pod fits any of the given nodes, based on
-// multiple criteria, like, pod node selector matching the node label, node
-// being schedulable or not.
-func PodFitsAnyNode(pod *v1.Pod, nodes []*v1.Node) bool {
-	for _, node := range nodes {
+// fitsRequest determines if a pod can fit on a node based on its resource requests. It returns true if
+// the pod will fit.
+func fitsRequest(nodeIndexer podutil.GetPodsAssignedToNodeFunc, pod *v1.Pod, node *v1.Node) (bool, []error) {
+	var insufficientResources []error
 
-		ok, err := utils.PodMatchNodeSelector(pod, node)
-		if err != nil || !ok {
-			continue
-		}
-		if !IsNodeUnschedulable(node) {
-			klog.V(2).InfoS("Pod can possibly be scheduled on a different node", "pod", klog.KObj(pod), "node", klog.KObj(node))
-			return true
-		}
+	// Get pod requests
+	podRequests, _ := utils.PodRequestsAndLimits(pod)
+	resourceNames := make([]v1.ResourceName, 0, len(podRequests))
+	for name := range podRequests {
+		resourceNames = append(resourceNames, name)
 	}
-	return false
-}
-
-// PodFitsCurrentNode checks if the given pod fits on the given node if the pod
-// node selector matches the node label.
-func PodFitsCurrentNode(pod *v1.Pod, node *v1.Node) bool {
-	ok, err := utils.PodMatchNodeSelector(pod, node)
 
+	availableResources, err := nodeAvailableResources(nodeIndexer, node, resourceNames)
 	if err != nil {
-		klog.ErrorS(err, "Failed to match node selector")
-		return false
+		return false, []error{err}
 	}
 
-	if !ok {
-		klog.V(2).InfoS("Pod does not fit on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
-		return false
+	podFitsOnNode := true
+	for _, resource := range resourceNames {
+		podResourceRequest := podRequests[resource]
+		availableResource, ok := availableResources[resource]
+		if !ok || podResourceRequest.MilliValue() > availableResource.MilliValue() {
+			insufficientResources = append(insufficientResources, fmt.Errorf("insufficient %v", resource))
+			podFitsOnNode = false
+		}
+	}
+	return podFitsOnNode, insufficientResources
+}
+
+// nodeAvailableResources returns resources mapped to the quanitity available on the node.
+func nodeAvailableResources(nodeIndexer podutil.GetPodsAssignedToNodeFunc, node *v1.Node, resourceNames []v1.ResourceName) (map[v1.ResourceName]*resource.Quantity, error) {
+	podsOnNode, err := podutil.ListPodsOnANode(node.Name, nodeIndexer, nil)
+	if err != nil {
+		return nil, err
+	}
+	nodeUtilization := NodeUtilization(podsOnNode, resourceNames)
+	remainingResources := map[v1.ResourceName]*resource.Quantity{
+		v1.ResourceCPU:    resource.NewMilliQuantity(node.Status.Allocatable.Cpu().MilliValue()-nodeUtilization[v1.ResourceCPU].MilliValue(), resource.DecimalSI),
+		v1.ResourceMemory: resource.NewQuantity(node.Status.Allocatable.Memory().Value()-nodeUtilization[v1.ResourceMemory].Value(), resource.BinarySI),
+		v1.ResourcePods:   resource.NewQuantity(node.Status.Allocatable.Pods().Value()-nodeUtilization[v1.ResourcePods].Value(), resource.DecimalSI),
+	}
+	for _, name := range resourceNames {
+		if !IsBasicResource(name) {
+			if _, exists := node.Status.Allocatable[name]; exists {
+				allocatableResource := node.Status.Allocatable[name]
+				remainingResources[name] = resource.NewQuantity(allocatableResource.Value()-nodeUtilization[name].Value(), resource.DecimalSI)
+			} else {
+				remainingResources[name] = resource.NewQuantity(0, resource.DecimalSI)
+			}
+		}
+	}
+
+	return remainingResources, nil
+}
+
+// NodeUtilization returns the resources requested by the given pods. Only resources supplied in the resourceNames parameter are calculated.
+func NodeUtilization(pods []*v1.Pod, resourceNames []v1.ResourceName) map[v1.ResourceName]*resource.Quantity {
+	totalReqs := map[v1.ResourceName]*resource.Quantity{
+		v1.ResourceCPU:    resource.NewMilliQuantity(0, resource.DecimalSI),
+		v1.ResourceMemory: resource.NewQuantity(0, resource.BinarySI),
+		v1.ResourcePods:   resource.NewQuantity(int64(len(pods)), resource.DecimalSI),
+	}
+	for _, name := range resourceNames {
+		if !IsBasicResource(name) {
+			totalReqs[name] = resource.NewQuantity(0, resource.DecimalSI)
+		}
+	}
+
+	for _, pod := range pods {
+		req, _ := utils.PodRequestsAndLimits(pod)
+		for _, name := range resourceNames {
+			quantity, ok := req[name]
+			if ok && name != v1.ResourcePods {
+				// As Quantity.Add says: Add adds the provided y quantity to the current value. If the current value is zero,
+				// the format of the quantity will be updated to the format of y.
+				totalReqs[name].Add(quantity)
+			}
+		}
+	}
+
+	return totalReqs
+}
+
+// IsBasicResource checks if resource is basic native.
+func IsBasicResource(name v1.ResourceName) bool {
+	switch name {
+	case v1.ResourceCPU, v1.ResourceMemory, v1.ResourcePods:
+		return true
+	default:
+		return false
 	}
-
-	klog.V(2).InfoS("Pod fits on node", "pod", klog.KObj(pod), "node", klog.KObj(node))
-	return true
 }

pkg/descheduler/node/node_test.go

@@ -21,9 +21,12 @@ import (
 	"testing"
 
 	v1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/runtime"
 	"k8s.io/client-go/informers"
 	"k8s.io/client-go/kubernetes/fake"
+	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
 	"sigs.k8s.io/descheduler/test"
 )
 
@@ -53,7 +56,6 @@ func TestReadyNodes(t *testing.T) {
 	if IsReady(node5) {
 		t.Errorf("Expected %v to be not ready", node5.Name)
 	}
-
 }
 
 func TestReadyNodesWithNodeSelector(t *testing.T) {
@@ -67,14 +69,14 @@ func TestReadyNodesWithNodeSelector(t *testing.T) {
 	nodeSelector := "type=compute"
 
 	sharedInformerFactory := informers.NewSharedInformerFactory(fakeClient, 0)
-	nodeInformer := sharedInformerFactory.Core().V1().Nodes()
+	nodeLister := sharedInformerFactory.Core().V1().Nodes().Lister()
 
 	stopChannel := make(chan struct{})
 	sharedInformerFactory.Start(stopChannel)
 	sharedInformerFactory.WaitForCacheSync(stopChannel)
 	defer close(stopChannel)
 
-	nodes, _ := ReadyNodes(ctx, fakeClient, nodeInformer, nodeSelector)
+	nodes, _ := ReadyNodes(ctx, fakeClient, nodeLister, nodeSelector)
 
 	if nodes[0].Name != "node1" {
 		t.Errorf("Expected node1, got %s", nodes[0].Name)
@@ -108,11 +110,9 @@ func TestIsNodeUnschedulable(t *testing.T) {
 			t.Errorf("Test %#v failed", test.description)
 		}
 	}
-
 }
 
 func TestPodFitsCurrentNode(t *testing.T) {
-
 	nodeLabelKey := "kubernetes.io/desiredNode"
 	nodeLabelValue := "yes"
 
@@ -147,13 +147,13 @@ func TestPodFitsCurrentNode(t *testing.T) {
 					},
 				},
 			},
-			node: &v1.Node{
-				ObjectMeta: metav1.ObjectMeta{
-					Labels: map[string]string{
-						nodeLabelKey: nodeLabelValue,
-					},
-				},
-			},
+			node: test.BuildTestNode("node1", 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+				node.ObjectMeta.Labels = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+
+				node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+			}),
 			success: true,
 		},
 		{
@@ -181,27 +181,48 @@ func TestPodFitsCurrentNode(t *testing.T) {
 					},
 				},
 			},
-			node: &v1.Node{
-				ObjectMeta: metav1.ObjectMeta{
-					Labels: map[string]string{
-						nodeLabelKey: "no",
-					},
-				},
-			},
+			node: test.BuildTestNode("node1", 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+				node.ObjectMeta.Labels = map[string]string{
+					nodeLabelKey: "no",
+				}
+
+				node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+			}),
 			success: false,
 		},
 	}
 
 	for _, tc := range tests {
-		actual := PodFitsCurrentNode(tc.pod, tc.node)
-		if actual != tc.success {
-			t.Errorf("Test %#v failed", tc.description)
-		}
+		t.Run(tc.description, func(t *testing.T) {
+			ctx, cancel := context.WithCancel(context.Background())
+			defer cancel()
+
+			var objs []runtime.Object
+			objs = append(objs, tc.node)
+			objs = append(objs, tc.pod)
+
+			fakeClient := fake.NewSimpleClientset(objs...)
+
+			sharedInformerFactory := informers.NewSharedInformerFactory(fakeClient, 0)
+			podInformer := sharedInformerFactory.Core().V1().Pods().Informer()
+
+			getPodsAssignedToNode, err := podutil.BuildGetPodsAssignedToNodeFunc(podInformer)
+			if err != nil {
+				t.Errorf("Build get pods assigned to node function error: %v", err)
+			}
+
+			sharedInformerFactory.Start(ctx.Done())
+			sharedInformerFactory.WaitForCacheSync(ctx.Done())
+
+			actual := PodFitsCurrentNode(getPodsAssignedToNode, tc.pod, tc.node)
+			if actual != tc.success {
+				t.Errorf("Test %#v failed", tc.description)
+			}
+		})
 	}
 }
 
 func TestPodFitsAnyOtherNode(t *testing.T) {
-
 	nodeLabelKey := "kubernetes.io/desiredNode"
 	nodeLabelValue := "yes"
 	nodeTaintKey := "hardware"
@@ -215,238 +236,527 @@ func TestPodFitsAnyOtherNode(t *testing.T) {
 		pod         *v1.Pod
 		nodes       []*v1.Node
 		success     bool
+		podsOnNodes []*v1.Pod
 	}{
 		{
 			description: "Pod fits another node matching node affinity",
-			pod:         createPodManifest(nodeNames[2], nodeLabelKey, nodeLabelValue),
+			pod: test.BuildTestPod("p1", 0, 0, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+			}),
 			nodes: []*v1.Node{
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[0],
-						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[1],
-						Labels: map[string]string{
-							nodeLabelKey: "no",
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[2],
-					},
-				},
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[1], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: "no",
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
 			},
-			success: true,
+			podsOnNodes: []*v1.Pod{},
+			success:     true,
 		},
 		{
 			description: "Pod expected to fit one of the nodes",
-			pod:         createPodManifest(nodeNames[2], nodeLabelKey, nodeLabelValue),
+			pod: test.BuildTestPod("p1", 0, 0, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+			}),
 			nodes: []*v1.Node{
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[0],
-						Labels: map[string]string{
-							nodeLabelKey: "no",
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[1],
-						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[2],
-					},
-				},
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: "no",
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[1], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
 			},
-			success: true,
+			podsOnNodes: []*v1.Pod{},
+			success:     true,
 		},
 		{
 			description: "Pod expected to fit none of the nodes",
-			pod:         createPodManifest(nodeNames[2], nodeLabelKey, nodeLabelValue),
+			pod: test.BuildTestPod("p1", 0, 0, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+			}),
 			nodes: []*v1.Node{
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[0],
-						Labels: map[string]string{
-							nodeLabelKey: "unfit1",
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[1],
-						Labels: map[string]string{
-							nodeLabelKey: "unfit2",
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[2],
-					},
-				},
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: "unfit1",
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[1], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: "unfit2",
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
 			},
-			success: false,
+			podsOnNodes: []*v1.Pod{},
+			success:     false,
 		},
 		{
 			description: "Nodes are unschedulable but labels match, should fail",
-			pod:         createPodManifest(nodeNames[2], nodeLabelKey, nodeLabelValue),
+			pod: test.BuildTestPod("p1", 0, 0, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+			}),
 			nodes: []*v1.Node{
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[0],
-						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
-						},
-					},
-					Spec: v1.NodeSpec{
-						Unschedulable: true,
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[1],
-						Labels: map[string]string{
-							nodeLabelKey: "no",
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[2],
-					},
-				},
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Unschedulable = true
+				}),
+				test.BuildTestNode(nodeNames[1], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: "no",
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
 			},
-			success: false,
+			podsOnNodes: []*v1.Pod{},
+			success:     false,
 		},
 		{
-			description: "Two nodes matches node selector, one of them is tained, should pass",
-			pod:         createPodManifest(nodeNames[2], nodeLabelKey, nodeLabelValue),
+			description: "Both nodes are tained, should fail",
+			pod: test.BuildTestPod("p1", 2000, 2*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
 			nodes: []*v1.Node{
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[0],
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+				}),
+				test.BuildTestNode(nodeNames[1], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{},
+			success:     false,
+		},
+		{
+			description: "Two nodes matches node selector, one of them is tained, there is a pod on the available node, and requests are low, should pass",
+			pod: test.BuildTestPod("p1", 2000, 2*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
+			nodes: []*v1.Node{
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+				}),
+				test.BuildTestNode(nodeNames[1], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("test-pod", 12*1000, 20*1000*1000*1000, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
 						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
+							"test": "true",
 						},
-					},
-					Spec: v1.NodeSpec{
-						Taints: []v1.Taint{
-							{
-								Key:    nodeTaintKey,
-								Value:  nodeTaintValue,
-								Effect: v1.TaintEffectNoSchedule,
-							},
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[1],
-						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
-						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[2],
-					},
-				},
+					}
+					pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(40*1000*1000*1000, resource.DecimalSI)
+				}),
 			},
 			success: true,
 		},
 		{
-			description: "Both nodes are tained, should fail",
-			pod:         createPodManifest(nodeNames[2], nodeLabelKey, nodeLabelValue),
+			description: "Two nodes matches node selector, one of them is tained, but CPU requests are too big, should fail",
+			pod: test.BuildTestPod("p1", 2000, 2*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
 			nodes: []*v1.Node{
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[0],
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+				}),
+				// Notice that this node only has 4 cores, the pod already on the node below requests 3 cores, and the pod above requests 2 cores
+				test.BuildTestNode(nodeNames[1], 4000, 8*1000*1000*1000, 12, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(200*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("3-core-pod", 3000, 4*1000*1000*1000, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
 						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
+							"test": "true",
 						},
-					},
-					Spec: v1.NodeSpec{
-						Taints: []v1.Taint{
-							{
-								Key:    nodeTaintKey,
-								Value:  nodeTaintValue,
-								Effect: v1.TaintEffectNoSchedule,
-							},
+					}
+					pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+				}),
+			},
+			success: false,
+		},
+		{
+			description: "Two nodes matches node selector, one of them is tained, but memory requests are too big, should fail",
+			pod: test.BuildTestPod("p1", 2000, 5*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
+			nodes: []*v1.Node{
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
 						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[1],
+					}
+				}),
+				// Notice that this node only has 8GB of memory, the pod already on the node below requests 4GB, and the pod above requests 5GB
+				test.BuildTestNode(nodeNames[1], 10*1000, 8*1000*1000*1000, 12, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(200*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("4GB-mem-pod", 2000, 4*1000*1000*1000, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
 						Labels: map[string]string{
-							nodeLabelKey: nodeLabelValue,
+							"test": "true",
 						},
-					},
-					Spec: v1.NodeSpec{
-						Taints: []v1.Taint{
-							{
-								Key:    nodeTaintKey,
-								Value:  nodeTaintValue,
-								Effect: v1.TaintEffectNoExecute,
-							},
+					}
+					pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+				}),
+			},
+			success: false,
+		},
+		{
+			description: "Two nodes matches node selector, one of them is tained, but ephemeral storage requests are too big, should fail",
+			pod: test.BuildTestPod("p1", 2000, 4*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
+			nodes: []*v1.Node{
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
 						},
-					},
-				},
-				{
-					ObjectMeta: metav1.ObjectMeta{
-						Name: nodeNames[2],
-					},
-				},
+					}
+				}),
+				// Notice that this node only has 20GB of storage, the pod already on the node below requests 11GB, and the pod above requests 10GB
+				test.BuildTestNode(nodeNames[1], 10*1000, 8*1000*1000*1000, 12, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(20*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("11GB-storage-pod", 2000, 4*1000*1000*1000, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
+						Labels: map[string]string{
+							"test": "true",
+						},
+					}
+					pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(11*1000*1000*1000, resource.DecimalSI)
+				}),
+			},
+			success: false,
+		},
+		{
+			description: "Two nodes matches node selector, one of them is tained, but custom resource requests are too big, should fail",
+			pod: test.BuildTestPod("p1", 2000, 2*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+				pod.Spec.Containers[0].Resources.Requests["example.com/custom-resource"] = *resource.NewQuantity(10, resource.DecimalSI)
+			}),
+			nodes: []*v1.Node{
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+					node.Status.Allocatable["example.com/custom-resource"] = *resource.NewQuantity(15, resource.DecimalSI)
+				}),
+				// Notice that this node only has 15 of the custom resource, the pod already on the node below requests 10, and the pod above requests 10
+				test.BuildTestNode(nodeNames[1], 10*1000, 8*1000*1000*1000, 12, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(200*1000*1000*1000, resource.DecimalSI)
+					node.Status.Allocatable["example.com/custom-resource"] = *resource.NewQuantity(15, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("10-custom-resource-pod", 0, 0, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
+						Labels: map[string]string{
+							"test": "true",
+						},
+					}
+					pod.Spec.Containers[0].Resources.Requests["example.com/custom-resource"] = *resource.NewQuantity(10, resource.DecimalSI)
+				}),
+			},
+			success: false,
+		},
+		{
+			description: "Two nodes matches node selector, one of them is tained, CPU requests will fit, and pod Overhead is low enough, should pass",
+			pod: test.BuildTestPod("p1", 1000, 2*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
+			nodes: []*v1.Node{
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+				}),
+				// Notice that this node has 5 CPU cores, the pod below requests 2 cores, and has CPU overhead of 1 cores, and the pod above requests 1 core
+				test.BuildTestNode(nodeNames[1], 5000, 8*1000*1000*1000, 12, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(200*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("3-core-pod", 2000, 4*1000*1000*1000, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
+						Labels: map[string]string{
+							"test": "true",
+						},
+					}
+					pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+					pod.Spec.Overhead = createResourceList(1000, 1000*1000*1000, 1000*1000*1000)
+				}),
+			},
+			success: true,
+		},
+		{
+			description: "Two nodes matches node selector, one of them is tained, CPU requests will fit, but pod Overhead is too high, should fail",
+			pod: test.BuildTestPod("p1", 2000, 2*1000*1000*1000, nodeNames[2], func(pod *v1.Pod) {
+				pod.Spec.NodeSelector = map[string]string{
+					nodeLabelKey: nodeLabelValue,
+				}
+				pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+			}),
+			nodes: []*v1.Node{
+				test.BuildTestNode(nodeNames[0], 64000, 128*1000*1000*1000, 200, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(1000*1000*1000*1000, resource.DecimalSI)
+					node.Spec.Taints = []v1.Taint{
+						{
+							Key:    nodeTaintKey,
+							Value:  nodeTaintValue,
+							Effect: v1.TaintEffectNoSchedule,
+						},
+					}
+				}),
+				// Notice that this node only has 5 CPU cores, the pod below requests 2 cores, but has CPU overhead of 2 cores, and the pod above requests 2 cores
+				test.BuildTestNode(nodeNames[1], 5000, 8*1000*1000*1000, 12, func(node *v1.Node) {
+					node.ObjectMeta.Labels = map[string]string{
+						nodeLabelKey: nodeLabelValue,
+					}
+
+					node.Status.Allocatable[v1.ResourceEphemeralStorage] = *resource.NewQuantity(200*1000*1000*1000, resource.DecimalSI)
+				}),
+				test.BuildTestNode(nodeNames[2], 0, 0, 0, nil),
+			},
+			podsOnNodes: []*v1.Pod{
+				test.BuildTestPod("3-core-pod", 2000, 4*1000*1000*1000, nodeNames[1], func(pod *v1.Pod) {
+					pod.ObjectMeta = metav1.ObjectMeta{
+						Namespace: "test",
+						Labels: map[string]string{
+							"test": "true",
+						},
+					}
+					pod.Spec.Containers[0].Resources.Requests[v1.ResourceEphemeralStorage] = *resource.NewQuantity(10*1000*1000*1000, resource.DecimalSI)
+					pod.Spec.Overhead = createResourceList(2000, 1000*1000*1000, 1000*1000*1000)
+				}),
 			},
 			success: false,
 		},
 	}
 
 	for _, tc := range tests {
-		actual := PodFitsAnyOtherNode(tc.pod, tc.nodes)
-		if actual != tc.success {
-			t.Errorf("Test %#v failed", tc.description)
-		}
+		t.Run(tc.description, func(t *testing.T) {
+			ctx, cancel := context.WithCancel(context.Background())
+			defer cancel()
+
+			var objs []runtime.Object
+			for _, node := range tc.nodes {
+				objs = append(objs, node)
+			}
+			for _, pod := range tc.podsOnNodes {
+				objs = append(objs, pod)
+			}
+			objs = append(objs, tc.pod)
+
+			fakeClient := fake.NewSimpleClientset(objs...)
+
+			sharedInformerFactory := informers.NewSharedInformerFactory(fakeClient, 0)
+			podInformer := sharedInformerFactory.Core().V1().Pods().Informer()
+
+			getPodsAssignedToNode, err := podutil.BuildGetPodsAssignedToNodeFunc(podInformer)
+			if err != nil {
+				t.Errorf("Build get pods assigned to node function error: %v", err)
+			}
+
+			sharedInformerFactory.Start(ctx.Done())
+			sharedInformerFactory.WaitForCacheSync(ctx.Done())
+
+			actual := PodFitsAnyOtherNode(getPodsAssignedToNode, tc.pod, tc.nodes)
+			if actual != tc.success {
+				t.Errorf("Test %#v failed", tc.description)
+			}
+		})
 	}
 }
 
-func createPodManifest(nodeName string, nodeSelectorKey string, nodeSelectorValue string) *v1.Pod {
-	return (&v1.Pod{
-		Spec: v1.PodSpec{
-			NodeName: nodeName,
-			Affinity: &v1.Affinity{
-				NodeAffinity: &v1.NodeAffinity{
-					RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
-						NodeSelectorTerms: []v1.NodeSelectorTerm{
-							{
-								MatchExpressions: []v1.NodeSelectorRequirement{
-									{
-										Key:      nodeSelectorKey,
-										Operator: "In",
-										Values: []string{
-											nodeSelectorValue,
-										},
-									},
-								},
-							},
-						},
-					},
-				},
-			},
-		},
-	})
+// createResourceList builds a small resource list of core resources
+func createResourceList(cpu, memory, ephemeralStorage int64) v1.ResourceList {
+	resourceList := make(map[v1.ResourceName]resource.Quantity)
+	resourceList[v1.ResourceCPU] = *resource.NewMilliQuantity(cpu, resource.DecimalSI)
+	resourceList[v1.ResourceMemory] = *resource.NewQuantity(memory, resource.DecimalSI)
+	resourceList[v1.ResourceEphemeralStorage] = *resource.NewQuantity(ephemeralStorage, resource.DecimalSI)
+	return resourceList
 }

pkg/descheduler/pod/pods.go

@@ -17,150 +17,173 @@ limitations under the License.
 package pod
 
 import (
-	"context"
 	"sort"
 
 	v1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/fields"
-	clientset "k8s.io/client-go/kubernetes"
+	"k8s.io/apimachinery/pkg/labels"
+	"k8s.io/apimachinery/pkg/util/sets"
+	"k8s.io/client-go/tools/cache"
+
 	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
+const (
+	nodeNameKeyIndex = "spec.nodeName"
+)
+
+// FilterFunc is a filter for a pod.
+type FilterFunc func(*v1.Pod) bool
+
+// GetPodsAssignedToNodeFunc is a function which accept a node name and a pod filter function
+// as input and returns the pods that assigned to the node.
+type GetPodsAssignedToNodeFunc func(string, FilterFunc) ([]*v1.Pod, error)
+
+// WrapFilterFuncs wraps a set of FilterFunc in one.
+func WrapFilterFuncs(filters ...FilterFunc) FilterFunc {
+	return func(pod *v1.Pod) bool {
+		for _, filter := range filters {
+			if filter != nil && !filter(pod) {
+				return false
+			}
+		}
+		return true
+	}
+}
+
 type Options struct {
-	filter             func(pod *v1.Pod) bool
-	includedNamespaces []string
-	excludedNamespaces []string
+	filter             FilterFunc
+	includedNamespaces sets.String
+	excludedNamespaces sets.String
 	labelSelector      *metav1.LabelSelector
 }
 
+// NewOptions returns an empty Options.
+func NewOptions() *Options {
+	return &Options{}
+}
+
 // WithFilter sets a pod filter.
 // The filter function should return true if the pod should be returned from ListPodsOnANode
-func WithFilter(filter func(pod *v1.Pod) bool) func(opts *Options) {
-	return func(opts *Options) {
-		opts.filter = filter
-	}
+func (o *Options) WithFilter(filter FilterFunc) *Options {
+	o.filter = filter
+	return o
 }
 
 // WithNamespaces sets included namespaces
-func WithNamespaces(namespaces []string) func(opts *Options) {
-	return func(opts *Options) {
-		opts.includedNamespaces = namespaces
-	}
+func (o *Options) WithNamespaces(namespaces sets.String) *Options {
+	o.includedNamespaces = namespaces
+	return o
 }
 
 // WithoutNamespaces sets excluded namespaces
-func WithoutNamespaces(namespaces []string) func(opts *Options) {
-	return func(opts *Options) {
-		opts.excludedNamespaces = namespaces
-	}
+func (o *Options) WithoutNamespaces(namespaces sets.String) *Options {
+	o.excludedNamespaces = namespaces
+	return o
 }
 
 // WithLabelSelector sets a pod label selector
-func WithLabelSelector(labelSelector *metav1.LabelSelector) func(opts *Options) {
-	return func(opts *Options) {
-		opts.labelSelector = labelSelector
-	}
+func (o *Options) WithLabelSelector(labelSelector *metav1.LabelSelector) *Options {
+	o.labelSelector = labelSelector
+	return o
 }
 
-// ListPodsOnANode lists all of the pods on a node
-// It also accepts an optional "filter" function which can be used to further limit the pods that are returned.
-// (Usually this is podEvictor.Evictable().IsEvictable, in order to only list the evictable pods on a node, but can
-// be used by strategies to extend it if there are further restrictions, such as with NodeAffinity).
-func ListPodsOnANode(
-	ctx context.Context,
-	client clientset.Interface,
-	node *v1.Node,
-	opts ...func(opts *Options),
-) ([]*v1.Pod, error) {
-	fieldSelectorString := "spec.nodeName=" + node.Name + ",status.phase!=" + string(v1.PodSucceeded) + ",status.phase!=" + string(v1.PodFailed)
-
-	return ListPodsOnANodeWithFieldSelector(ctx, client, node, fieldSelectorString, opts...)
+// BuildFilterFunc builds a final FilterFunc based on Options.
+func (o *Options) BuildFilterFunc() (FilterFunc, error) {
+	var s labels.Selector
+	var err error
+	if o.labelSelector != nil {
+		s, err = metav1.LabelSelectorAsSelector(o.labelSelector)
+		if err != nil {
+			return nil, err
+		}
+	}
+	return func(pod *v1.Pod) bool {
+		if o.filter != nil && !o.filter(pod) {
+			return false
+		}
+		if len(o.includedNamespaces) > 0 && !o.includedNamespaces.Has(pod.Namespace) {
+			return false
+		}
+		if len(o.excludedNamespaces) > 0 && o.excludedNamespaces.Has(pod.Namespace) {
+			return false
+		}
+		if s != nil && !s.Matches(labels.Set(pod.GetLabels())) {
+			return false
+		}
+		return true
+	}, nil
 }
 
-// ListPodsOnANodeWithFieldSelector lists all of the pods on a node using the filter selectors
-func ListPodsOnANodeWithFieldSelector(
-	ctx context.Context,
-	client clientset.Interface,
-	node *v1.Node,
-	fieldSelectorString string,
-	opts ...func(opts *Options),
-) ([]*v1.Pod, error) {
-	options := &Options{}
-	for _, opt := range opts {
-		opt(options)
-	}
-
-	pods := make([]*v1.Pod, 0)
-
-	labelSelectorString := ""
-	if options.labelSelector != nil {
-		selector, err := metav1.LabelSelectorAsSelector(options.labelSelector)
-		if err != nil {
-			return []*v1.Pod{}, err
-		}
-		labelSelectorString = selector.String()
-	}
-
-	if len(options.includedNamespaces) > 0 {
-		fieldSelector, err := fields.ParseSelector(fieldSelectorString)
-		if err != nil {
-			return []*v1.Pod{}, err
-		}
-
-		for _, namespace := range options.includedNamespaces {
-			podList, err := client.CoreV1().Pods(namespace).List(ctx,
-				metav1.ListOptions{
-					FieldSelector: fieldSelector.String(),
-					LabelSelector: labelSelectorString,
-				})
-			if err != nil {
-				return []*v1.Pod{}, err
+// BuildGetPodsAssignedToNodeFunc establishes an indexer to map the pods and their assigned nodes.
+// It returns a function to help us get all the pods that assigned to a node based on the indexer.
+func BuildGetPodsAssignedToNodeFunc(podInformer cache.SharedIndexInformer) (GetPodsAssignedToNodeFunc, error) {
+	// Establish an indexer to map the pods and their assigned nodes.
+	err := podInformer.AddIndexers(cache.Indexers{
+		nodeNameKeyIndex: func(obj interface{}) ([]string, error) {
+			pod, ok := obj.(*v1.Pod)
+			if !ok {
+				return []string{}, nil
 			}
-			for i := range podList.Items {
-				if options.filter != nil && !options.filter(&podList.Items[i]) {
-					continue
-				}
-				pods = append(pods, &podList.Items[i])
+			if len(pod.Spec.NodeName) == 0 {
+				return []string{}, nil
+			}
+			return []string{pod.Spec.NodeName}, nil
+		},
+	})
+	if err != nil {
+		return nil, err
+	}
+
+	// The indexer helps us get all the pods that assigned to a node.
+	podIndexer := podInformer.GetIndexer()
+	getPodsAssignedToNode := func(nodeName string, filter FilterFunc) ([]*v1.Pod, error) {
+		objs, err := podIndexer.ByIndex(nodeNameKeyIndex, nodeName)
+		if err != nil {
+			return nil, err
+		}
+		pods := make([]*v1.Pod, 0, len(objs))
+		for _, obj := range objs {
+			pod, ok := obj.(*v1.Pod)
+			if !ok {
+				continue
+			}
+			if filter(pod) {
+				pods = append(pods, pod)
 			}
 		}
 		return pods, nil
 	}
+	return getPodsAssignedToNode, nil
+}
 
-	if len(options.excludedNamespaces) > 0 {
-		for _, namespace := range options.excludedNamespaces {
-			fieldSelectorString += ",metadata.namespace!=" + namespace
-		}
+// ListPodsOnANode lists all pods on a node.
+// It also accepts a "filter" function which can be used to further limit the pods that are returned.
+// (Usually this is podEvictor.Evictable().IsEvictable, in order to only list the evictable pods on a node, but can
+// be used by strategies to extend it if there are further restrictions, such as with NodeAffinity).
+func ListPodsOnANode(
+	nodeName string,
+	getPodsAssignedToNode GetPodsAssignedToNodeFunc,
+	filter FilterFunc,
+) ([]*v1.Pod, error) {
+	// Succeeded and failed pods are not considered because they don't occupy any resource.
+	f := func(pod *v1.Pod) bool {
+		return pod.Status.Phase != v1.PodSucceeded && pod.Status.Phase != v1.PodFailed
 	}
+	return ListAllPodsOnANode(nodeName, getPodsAssignedToNode, WrapFilterFuncs(f, filter))
+}
 
-	fieldSelector, err := fields.ParseSelector(fieldSelectorString)
+// ListAllPodsOnANode lists all the pods on a node no matter what the phase of the pod is.
+func ListAllPodsOnANode(
+	nodeName string,
+	getPodsAssignedToNode GetPodsAssignedToNodeFunc,
+	filter FilterFunc,
+) ([]*v1.Pod, error) {
+	pods, err := getPodsAssignedToNode(nodeName, filter)
 	if err != nil {
 		return []*v1.Pod{}, err
 	}
 
-	// INFO(jchaloup): field selectors do not work properly with listers
-	// Once the descheduler switches to pod listers (through informers),
-	// We need to flip to client-side filtering.
-	podList, err := client.CoreV1().Pods(v1.NamespaceAll).List(ctx,
-		metav1.ListOptions{
-			FieldSelector: fieldSelector.String(),
-			LabelSelector: labelSelectorString,
-		})
-	if err != nil {
-		return []*v1.Pod{}, err
-	}
-
-	for i := range podList.Items {
-		// fake client does not support field selectors
-		// so let's filter based on the node name as well (quite cheap)
-		if podList.Items[i].Spec.NodeName != node.Name {
-			continue
-		}
-		if options.filter != nil && !options.filter(&podList.Items[i]) {
-			continue
-		}
-		pods = append(pods, &podList.Items[i])
-	}
 	return pods, nil
 }
 
@@ -204,3 +227,10 @@ func SortPodsBasedOnPriorityLowToHigh(pods []*v1.Pod) {
 		return *pods[i].Spec.Priority < *pods[j].Spec.Priority
 	})
 }
+
+// SortPodsBasedOnAge sorts Pods from oldest to most recent in place
+func SortPodsBasedOnAge(pods []*v1.Pod) {
+	sort.Slice(pods, func(i, j int) bool {
+		return pods[i].CreationTimestamp.Before(&pods[j].CreationTimestamp)
+	})
+}

pkg/descheduler/pod/pods_test.go

@@ -20,14 +20,15 @@ import (
 	"context"
 	"fmt"
 	"reflect"
-	"strings"
 	"testing"
+	"time"
 
 	v1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/client-go/informers"
 	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
+
 	"sigs.k8s.io/descheduler/test"
 )
 
@@ -39,19 +40,17 @@ var (
 func TestListPodsOnANode(t *testing.T) {
 	testCases := []struct {
 		name             string
-		pods             map[string][]v1.Pod
+		pods             []*v1.Pod
 		node             *v1.Node
 		labelSelector    *metav1.LabelSelector
 		expectedPodCount int
 	}{
 		{
 			name: "test listing pods on a node",
-			pods: map[string][]v1.Pod{
-				"n1": {
-					*test.BuildTestPod("pod1", 100, 0, "n1", nil),
-					*test.BuildTestPod("pod2", 100, 0, "n1", nil),
-				},
-				"n2": {*test.BuildTestPod("pod3", 100, 0, "n2", nil)},
+			pods: []*v1.Pod{
+				test.BuildTestPod("pod1", 100, 0, "n1", nil),
+				test.BuildTestPod("pod2", 100, 0, "n1", nil),
+				test.BuildTestPod("pod3", 100, 0, "n2", nil),
 			},
 			node:             test.BuildTestNode("n1", 2000, 3000, 10, nil),
 			labelSelector:    nil,
@@ -59,17 +58,15 @@ func TestListPodsOnANode(t *testing.T) {
 		},
 		{
 			name: "test listing pods with label selector",
-			pods: map[string][]v1.Pod{
-				"n1": {
-					*test.BuildTestPod("pod1", 100, 0, "n1", nil),
-					*test.BuildTestPod("pod2", 100, 0, "n1", func(pod *v1.Pod) {
-						pod.Labels = map[string]string{"foo": "bar"}
-					}),
-					*test.BuildTestPod("pod3", 100, 0, "n1", func(pod *v1.Pod) {
-						pod.Labels = map[string]string{"foo": "bar1"}
-					}),
-				},
-				"n2": {*test.BuildTestPod("pod4", 100, 0, "n2", nil)},
+			pods: []*v1.Pod{
+				test.BuildTestPod("pod1", 100, 0, "n1", nil),
+				test.BuildTestPod("pod2", 100, 0, "n1", func(pod *v1.Pod) {
+					pod.Labels = map[string]string{"foo": "bar"}
+				}),
+				test.BuildTestPod("pod3", 100, 0, "n1", func(pod *v1.Pod) {
+					pod.Labels = map[string]string{"foo": "bar1"}
+				}),
+				test.BuildTestPod("pod4", 100, 0, "n2", nil),
 			},
 			node: test.BuildTestNode("n1", 2000, 3000, 10, nil),
 			labelSelector: &metav1.LabelSelector{
@@ -85,21 +82,38 @@ func TestListPodsOnANode(t *testing.T) {
 		},
 	}
 	for _, testCase := range testCases {
-		fakeClient := &fake.Clientset{}
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			list := action.(core.ListAction)
-			fieldString := list.GetListRestrictions().Fields.String()
-			if strings.Contains(fieldString, "n1") {
-				return true, &v1.PodList{Items: testCase.pods["n1"]}, nil
-			} else if strings.Contains(fieldString, "n2") {
-				return true, &v1.PodList{Items: testCase.pods["n2"]}, nil
+		t.Run(testCase.name, func(t *testing.T) {
+			ctx, cancel := context.WithCancel(context.Background())
+			defer cancel()
+
+			var objs []runtime.Object
+			objs = append(objs, testCase.node)
+			for _, pod := range testCase.pods {
+				objs = append(objs, pod)
+			}
+			fakeClient := fake.NewSimpleClientset(objs...)
+
+			sharedInformerFactory := informers.NewSharedInformerFactory(fakeClient, 0)
+			podInformer := sharedInformerFactory.Core().V1().Pods().Informer()
+
+			getPodsAssignedToNode, err := BuildGetPodsAssignedToNodeFunc(podInformer)
+			if err != nil {
+				t.Errorf("Build get pods assigned to node function error: %v", err)
+			}
+
+			sharedInformerFactory.Start(ctx.Done())
+			sharedInformerFactory.WaitForCacheSync(ctx.Done())
+
+			filter, err := NewOptions().WithLabelSelector(testCase.labelSelector).BuildFilterFunc()
+			if err != nil {
+				t.Errorf("Build filter function error: %v", err)
+			}
+
+			pods, _ := ListPodsOnANode(testCase.node.Name, getPodsAssignedToNode, filter)
+			if len(pods) != testCase.expectedPodCount {
+				t.Errorf("Expected %v pods on node %v, got %+v", testCase.expectedPodCount, testCase.node.Name, len(pods))
 			}
-			return true, nil, fmt.Errorf("Failed to list: %v", list)
 		})
-		pods, _ := ListPodsOnANode(context.TODO(), fakeClient, testCase.node, WithLabelSelector(testCase.labelSelector))
-		if len(pods) != testCase.expectedPodCount {
-			t.Errorf("expected %v pods on node %v, got %+v", testCase.expectedPodCount, testCase.node.Name, len(pods))
-		}
 	}
 }
 
@@ -142,3 +156,23 @@ func TestSortPodsBasedOnPriorityLowToHigh(t *testing.T) {
 		t.Errorf("Expected last pod in sorted list to be %v which of highest priority and guaranteed but got %v", p4, podList[len(podList)-1])
 	}
 }
+
+func TestSortPodsBasedOnAge(t *testing.T) {
+	podList := make([]*v1.Pod, 9)
+	n1 := test.BuildTestNode("n1", 4000, 3000, int64(len(podList)), nil)
+
+	for i := 0; i < len(podList); i++ {
+		podList[i] = test.BuildTestPod(fmt.Sprintf("p%d", i), 1, 32, n1.Name, func(pod *v1.Pod) {
+			creationTimestamp := metav1.Now().Add(time.Minute * time.Duration(-i))
+			pod.ObjectMeta.SetCreationTimestamp(metav1.NewTime(creationTimestamp))
+		})
+	}
+
+	SortPodsBasedOnAge(podList)
+
+	for i := 0; i < len(podList)-1; i++ {
+		if podList[i+1].CreationTimestamp.Before(&podList[i].CreationTimestamp) {
+			t.Errorf("Expected pods to be sorted by age but pod at index %d was older than %d", i+1, i)
+		}
+	}
+}

pkg/descheduler/policyconfig.go

@@ -17,24 +17,30 @@ limitations under the License.
 package descheduler
 
 import (
+	"context"
 	"fmt"
-	"io/ioutil"
+	"os"
 
 	"k8s.io/apimachinery/pkg/runtime"
+	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog/v2"
 
 	"sigs.k8s.io/descheduler/pkg/api"
 	"sigs.k8s.io/descheduler/pkg/api/v1alpha1"
 	"sigs.k8s.io/descheduler/pkg/descheduler/scheme"
+	"sigs.k8s.io/descheduler/pkg/framework"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/defaultevictor"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/pluginbuilder"
+	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
-func LoadPolicyConfig(policyConfigFile string) (*api.DeschedulerPolicy, error) {
+func LoadPolicyConfig(policyConfigFile string, client clientset.Interface, registry pluginbuilder.Registry) (*api.DeschedulerPolicy, error) {
 	if policyConfigFile == "" {
 		klog.V(1).InfoS("Policy config file not specified")
 		return nil, nil
 	}
 
-	policy, err := ioutil.ReadFile(policyConfigFile)
+	policy, err := os.ReadFile(policyConfigFile)
 	if err != nil {
 		return nil, fmt.Errorf("failed to read policy config file %q: %+v", policyConfigFile, err)
 	}
@@ -46,10 +52,164 @@ func LoadPolicyConfig(policyConfigFile string) (*api.DeschedulerPolicy, error) {
 		return nil, fmt.Errorf("failed decoding descheduler's policy config %q: %v", policyConfigFile, err)
 	}
 
-	internalPolicy := &api.DeschedulerPolicy{}
-	if err := scheme.Scheme.Convert(versionedPolicy, internalPolicy, nil); err != nil {
+	// Build profiles
+	internalPolicy, err := V1alpha1ToInternal(client, versionedPolicy, registry)
+	if err != nil {
 		return nil, fmt.Errorf("failed converting versioned policy to internal policy version: %v", err)
 	}
 
 	return internalPolicy, nil
 }
+
+func V1alpha1ToInternal(
+	client clientset.Interface,
+	deschedulerPolicy *v1alpha1.DeschedulerPolicy,
+	registry pluginbuilder.Registry,
+) (*api.DeschedulerPolicy, error) {
+	var evictLocalStoragePods bool
+	if deschedulerPolicy.EvictLocalStoragePods != nil {
+		evictLocalStoragePods = *deschedulerPolicy.EvictLocalStoragePods
+	}
+
+	evictBarePods := false
+	if deschedulerPolicy.EvictFailedBarePods != nil {
+		evictBarePods = *deschedulerPolicy.EvictFailedBarePods
+		if evictBarePods {
+			klog.V(1).InfoS("Warning: EvictFailedBarePods is set to True. This could cause eviction of pods without ownerReferences.")
+		}
+	}
+
+	evictSystemCriticalPods := false
+	if deschedulerPolicy.EvictSystemCriticalPods != nil {
+		evictSystemCriticalPods = *deschedulerPolicy.EvictSystemCriticalPods
+		if evictSystemCriticalPods {
+			klog.V(1).InfoS("Warning: EvictSystemCriticalPods is set to True. This could cause eviction of Kubernetes system pods.")
+		}
+	}
+
+	ignorePvcPods := false
+	if deschedulerPolicy.IgnorePVCPods != nil {
+		ignorePvcPods = *deschedulerPolicy.IgnorePVCPods
+	}
+
+	var profiles []api.Profile
+
+	// Build profiles
+	for name, strategy := range deschedulerPolicy.Strategies {
+		if _, ok := pluginbuilder.PluginRegistry[string(name)]; ok {
+			if strategy.Enabled {
+				params := strategy.Params
+				if params == nil {
+					params = &v1alpha1.StrategyParameters{}
+				}
+
+				nodeFit := false
+				if name != "PodLifeTime" {
+					nodeFit = params.NodeFit
+				}
+
+				// TODO(jchaloup): once all strategies are migrated move this check under
+				// the default evictor args validation
+				if params.ThresholdPriority != nil && params.ThresholdPriorityClassName != "" {
+					klog.ErrorS(fmt.Errorf("priority threshold misconfigured"), "only one of priorityThreshold fields can be set", "pluginName", name)
+					return nil, fmt.Errorf("priority threshold misconfigured for plugin %v", name)
+				}
+				var priorityThreshold *api.PriorityThreshold
+				if strategy.Params != nil {
+					priorityThreshold = &api.PriorityThreshold{
+						Value: strategy.Params.ThresholdPriority,
+						Name:  strategy.Params.ThresholdPriorityClassName,
+					}
+				}
+				thresholdPriority, err := utils.GetPriorityFromStrategyParams(context.TODO(), client, priorityThreshold)
+				if err != nil {
+					klog.ErrorS(err, "Failed to get threshold priority from strategy's params")
+					return nil, fmt.Errorf("failed to get threshold priority from strategy's params: %v", err)
+				}
+
+				var pluginConfig *api.PluginConfig
+				if pcFnc, exists := strategyParamsToPluginArgs[string(name)]; exists {
+					pluginConfig, err = pcFnc(params)
+					if err != nil {
+						klog.ErrorS(err, "skipping strategy", "strategy", name)
+						return nil, fmt.Errorf("failed to get plugin config for strategy %v: %v", name, err)
+					}
+				} else {
+					klog.ErrorS(fmt.Errorf("unknown strategy name"), "skipping strategy", "strategy", name)
+					return nil, fmt.Errorf("unknown strategy name: %v", name)
+				}
+
+				profile := api.Profile{
+					Name: fmt.Sprintf("strategy-%v-profile", name),
+					PluginConfigs: []api.PluginConfig{
+						{
+							Name: defaultevictor.PluginName,
+							Args: &defaultevictor.DefaultEvictorArgs{
+								EvictLocalStoragePods:   evictLocalStoragePods,
+								EvictSystemCriticalPods: evictSystemCriticalPods,
+								IgnorePvcPods:           ignorePvcPods,
+								EvictFailedBarePods:     evictBarePods,
+								NodeFit:                 nodeFit,
+								PriorityThreshold: &api.PriorityThreshold{
+									Value: &thresholdPriority,
+								},
+							},
+						},
+						*pluginConfig,
+					},
+					Plugins: api.Plugins{
+						Evict: api.PluginSet{
+							Enabled: []string{defaultevictor.PluginName},
+						},
+					},
+				}
+
+				pluginArgs := registry[string(name)].PluginArgInstance
+				pluginInstance, err := registry[string(name)].PluginBuilder(pluginArgs, &handleImpl{})
+				if err != nil {
+					klog.ErrorS(fmt.Errorf("could not build plugin"), "plugin build error", "plugin", name)
+					return nil, fmt.Errorf("could not build plugin: %v", name)
+				}
+
+				// pluginInstance can be of any of each type, or both
+				profilePlugins := profile.Plugins
+				profile.Plugins = enableProfilePluginsByType(profilePlugins, pluginInstance, pluginConfig)
+				profiles = append(profiles, profile)
+			}
+		} else {
+			klog.ErrorS(fmt.Errorf("unknown strategy name"), "skipping strategy", "strategy", name)
+			return nil, fmt.Errorf("unknown strategy name: %v", name)
+		}
+	}
+
+	return &api.DeschedulerPolicy{
+		Profiles:                       profiles,
+		NodeSelector:                   deschedulerPolicy.NodeSelector,
+		MaxNoOfPodsToEvictPerNode:      deschedulerPolicy.MaxNoOfPodsToEvictPerNode,
+		MaxNoOfPodsToEvictPerNamespace: deschedulerPolicy.MaxNoOfPodsToEvictPerNamespace,
+	}, nil
+}
+
+func enableProfilePluginsByType(profilePlugins api.Plugins, pluginInstance framework.Plugin, pluginConfig *api.PluginConfig) api.Plugins {
+	profilePlugins = checkBalance(profilePlugins, pluginInstance, pluginConfig)
+	profilePlugins = checkDeschedule(profilePlugins, pluginInstance, pluginConfig)
+	return profilePlugins
+}
+
+func checkBalance(profilePlugins api.Plugins, pluginInstance framework.Plugin, pluginConfig *api.PluginConfig) api.Plugins {
+	switch p := pluginInstance.(type) {
+	case framework.BalancePlugin:
+		klog.V(3).Info("converting Balance plugin: %s", p.Name())
+		profilePlugins.Balance.Enabled = []string{pluginConfig.Name}
+	}
+	return profilePlugins
+}
+
+func checkDeschedule(profilePlugins api.Plugins, pluginInstance framework.Plugin, pluginConfig *api.PluginConfig) api.Plugins {
+	switch p := pluginInstance.(type) {
+	case framework.DeschedulePlugin:
+		klog.V(3).Info("converting Deschedule plugin: %s", p.Name())
+		profilePlugins.Deschedule.Enabled = []string{pluginConfig.Name}
+	}
+	return profilePlugins
+}

pkg/descheduler/policyconfig_test.go

@@ -0,0 +1,701 @@
+/*
+Copyright 2017 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package descheduler
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/google/go-cmp/cmp"
+	fakeclientset "k8s.io/client-go/kubernetes/fake"
+	utilpointer "k8s.io/utils/pointer"
+	"sigs.k8s.io/descheduler/pkg/api"
+	"sigs.k8s.io/descheduler/pkg/api/v1alpha1"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/defaultevictor"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/nodeutilization"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/pluginbuilder"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removeduplicates"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removefailedpods"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodshavingtoomanyrestarts"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatinginterpodantiaffinity"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingnodeaffinity"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingnodetaints"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingtopologyspreadconstraint"
+	"sigs.k8s.io/descheduler/pkg/utils"
+)
+
+func TestV1alpha1ToV1alpha2(t *testing.T) {
+	SetupPlugins()
+	defaultEvictorPluginConfig := api.PluginConfig{
+		Name: defaultevictor.PluginName,
+		Args: &defaultevictor.DefaultEvictorArgs{
+			PriorityThreshold: &api.PriorityThreshold{
+				Value: utilpointer.Int32(utils.SystemCriticalPriority),
+			},
+		},
+	}
+	defaultEvictorPluginSet := api.PluginSet{
+		Enabled: []string{defaultevictor.PluginName},
+	}
+	type testCase struct {
+		description string
+		policy      *v1alpha1.DeschedulerPolicy
+		err         error
+		result      *api.DeschedulerPolicy
+	}
+	testCases := []testCase{
+		{
+			description: "RemoveFailedPods enabled, LowNodeUtilization disabled strategies to profile",
+			policy: &v1alpha1.DeschedulerPolicy{
+				Strategies: v1alpha1.StrategyList{
+					removeduplicates.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							Namespaces: &v1alpha1.Namespaces{
+								Exclude: []string{
+									"test2",
+								},
+							},
+						},
+					},
+					nodeutilization.LowNodeUtilizationPluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: false,
+						Params: &v1alpha1.StrategyParameters{
+							NodeResourceUtilizationThresholds: &v1alpha1.NodeResourceUtilizationThresholds{
+								Thresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(20),
+									"memory": v1alpha1.Percentage(20),
+									"pods":   v1alpha1.Percentage(20),
+								},
+								TargetThresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(50),
+									"memory": v1alpha1.Percentage(50),
+									"pods":   v1alpha1.Percentage(50),
+								},
+							},
+						},
+					},
+				},
+			},
+			result: &api.DeschedulerPolicy{
+				Profiles: []api.Profile{
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removeduplicates.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removeduplicates.PluginName,
+								Args: &removeduplicates.RemoveDuplicatesArgs{
+									Namespaces: &api.Namespaces{
+										Exclude: []string{
+											"test2",
+										},
+									},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{removeduplicates.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					// Disabled strategy is not generating internal plugin since it is not being used internally currently
+					// {
+					// 	Name: nodeutilization.LowNodeUtilizationPluginName,
+					// 	PluginConfigs: []api.PluginConfig{
+					// 		{
+					// 			Name: nodeutilization.LowNodeUtilizationPluginName,
+					// 			Args: &nodeutilization.LowNodeUtilizationArgs{
+					// 				Thresholds: api.ResourceThresholds{
+					// 					"cpu":    api.Percentage(20),
+					// [...]
+					// [...]
+					// },
+				},
+			},
+		},
+		{
+			description: "convert all strategies",
+			policy: &v1alpha1.DeschedulerPolicy{
+				Strategies: v1alpha1.StrategyList{
+					removeduplicates.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params:  &v1alpha1.StrategyParameters{},
+					},
+					nodeutilization.LowNodeUtilizationPluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							NodeResourceUtilizationThresholds: &v1alpha1.NodeResourceUtilizationThresholds{
+								Thresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(20),
+									"memory": v1alpha1.Percentage(20),
+									"pods":   v1alpha1.Percentage(20),
+								},
+								TargetThresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(50),
+									"memory": v1alpha1.Percentage(50),
+									"pods":   v1alpha1.Percentage(50),
+								},
+							},
+						},
+					},
+					nodeutilization.HighNodeUtilizationPluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							NodeResourceUtilizationThresholds: &v1alpha1.NodeResourceUtilizationThresholds{
+								Thresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(20),
+									"memory": v1alpha1.Percentage(20),
+									"pods":   v1alpha1.Percentage(20),
+								},
+							},
+						},
+					},
+					removefailedpods.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params:  &v1alpha1.StrategyParameters{},
+					},
+					removepodshavingtoomanyrestarts.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							PodsHavingTooManyRestarts: &v1alpha1.PodsHavingTooManyRestarts{
+								PodRestartThreshold: 100,
+							},
+						},
+					},
+					removepodsviolatinginterpodantiaffinity.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params:  &v1alpha1.StrategyParameters{},
+					},
+					removepodsviolatingnodeaffinity.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							NodeAffinityType: []string{"requiredDuringSchedulingIgnoredDuringExecution"},
+						},
+					},
+					removepodsviolatingnodetaints.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params:  &v1alpha1.StrategyParameters{},
+					},
+					removepodsviolatingtopologyspreadconstraint.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params:  &v1alpha1.StrategyParameters{},
+					},
+				},
+			},
+			result: &api.DeschedulerPolicy{
+				Profiles: []api.Profile{
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", nodeutilization.HighNodeUtilizationPluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: nodeutilization.HighNodeUtilizationPluginName,
+								Args: &nodeutilization.HighNodeUtilizationArgs{
+									Thresholds: api.ResourceThresholds{
+										"cpu":    api.Percentage(20),
+										"memory": api.Percentage(20),
+										"pods":   api.Percentage(20),
+									},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{nodeutilization.HighNodeUtilizationPluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", nodeutilization.LowNodeUtilizationPluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: nodeutilization.LowNodeUtilizationPluginName,
+								Args: &nodeutilization.LowNodeUtilizationArgs{
+									Thresholds: api.ResourceThresholds{
+										"cpu":    api.Percentage(20),
+										"memory": api.Percentage(20),
+										"pods":   api.Percentage(20),
+									},
+									TargetThresholds: api.ResourceThresholds{
+										"cpu":    api.Percentage(50),
+										"memory": api.Percentage(50),
+										"pods":   api.Percentage(50),
+									},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{nodeutilization.LowNodeUtilizationPluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removeduplicates.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removeduplicates.PluginName,
+								Args: &removeduplicates.RemoveDuplicatesArgs{},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{removeduplicates.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removefailedpods.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removefailedpods.PluginName,
+								Args: &removefailedpods.RemoveFailedPodsArgs{},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removefailedpods.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodshavingtoomanyrestarts.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodshavingtoomanyrestarts.PluginName,
+								Args: &removepodshavingtoomanyrestarts.RemovePodsHavingTooManyRestartsArgs{
+									PodRestartThreshold: 100,
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodshavingtoomanyrestarts.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatinginterpodantiaffinity.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatinginterpodantiaffinity.PluginName,
+								Args: &removepodsviolatinginterpodantiaffinity.RemovePodsViolatingInterPodAntiAffinityArgs{},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodsviolatinginterpodantiaffinity.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatingnodeaffinity.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatingnodeaffinity.PluginName,
+								Args: &removepodsviolatingnodeaffinity.RemovePodsViolatingNodeAffinityArgs{
+									NodeAffinityType: []string{"requiredDuringSchedulingIgnoredDuringExecution"},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodsviolatingnodeaffinity.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatingnodetaints.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatingnodetaints.PluginName,
+								Args: &removepodsviolatingnodetaints.RemovePodsViolatingNodeTaintsArgs{},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodsviolatingnodetaints.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatingtopologyspreadconstraint.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatingtopologyspreadconstraint.PluginName,
+								Args: &removepodsviolatingtopologyspreadconstraint.RemovePodsViolatingTopologySpreadConstraintArgs{},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{removepodsviolatingtopologyspreadconstraint.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+				},
+			},
+		},
+		{
+			description: "pass in all params to check args",
+			policy: &v1alpha1.DeschedulerPolicy{
+				Strategies: v1alpha1.StrategyList{
+					removeduplicates.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							RemoveDuplicates: &v1alpha1.RemoveDuplicates{
+								ExcludeOwnerKinds: []string{"ReplicaSet"},
+							},
+						},
+					},
+					nodeutilization.LowNodeUtilizationPluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							NodeResourceUtilizationThresholds: &v1alpha1.NodeResourceUtilizationThresholds{
+								Thresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(20),
+									"memory": v1alpha1.Percentage(20),
+									"pods":   v1alpha1.Percentage(20),
+								},
+								TargetThresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(50),
+									"memory": v1alpha1.Percentage(50),
+									"pods":   v1alpha1.Percentage(50),
+								},
+								UseDeviationThresholds: true,
+								NumberOfNodes:          3,
+							},
+						},
+					},
+					nodeutilization.HighNodeUtilizationPluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							NodeResourceUtilizationThresholds: &v1alpha1.NodeResourceUtilizationThresholds{
+								Thresholds: v1alpha1.ResourceThresholds{
+									"cpu":    v1alpha1.Percentage(20),
+									"memory": v1alpha1.Percentage(20),
+									"pods":   v1alpha1.Percentage(20),
+								},
+							},
+						},
+					},
+					removefailedpods.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							FailedPods: &v1alpha1.FailedPods{
+								MinPodLifetimeSeconds:   utilpointer.Uint(3600),
+								ExcludeOwnerKinds:       []string{"Job"},
+								Reasons:                 []string{"NodeAffinity"},
+								IncludingInitContainers: true,
+							},
+						},
+					},
+					removepodshavingtoomanyrestarts.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							PodsHavingTooManyRestarts: &v1alpha1.PodsHavingTooManyRestarts{
+								PodRestartThreshold:     100,
+								IncludingInitContainers: true,
+							},
+						},
+					},
+					removepodsviolatinginterpodantiaffinity.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params:  &v1alpha1.StrategyParameters{},
+					},
+					removepodsviolatingnodeaffinity.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							NodeAffinityType: []string{"requiredDuringSchedulingIgnoredDuringExecution"},
+						},
+					},
+					removepodsviolatingnodetaints.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							ExcludedTaints: []string{"dedicated=special-user", "reserved"},
+						},
+					},
+					removepodsviolatingtopologyspreadconstraint.PluginName: v1alpha1.DeschedulerStrategy{
+						Enabled: true,
+						Params: &v1alpha1.StrategyParameters{
+							IncludeSoftConstraints: true,
+						},
+					},
+				},
+			},
+			result: &api.DeschedulerPolicy{
+				Profiles: []api.Profile{
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", nodeutilization.HighNodeUtilizationPluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: nodeutilization.HighNodeUtilizationPluginName,
+								Args: &nodeutilization.HighNodeUtilizationArgs{
+									Thresholds: api.ResourceThresholds{
+										"cpu":    api.Percentage(20),
+										"memory": api.Percentage(20),
+										"pods":   api.Percentage(20),
+									},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{nodeutilization.HighNodeUtilizationPluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", nodeutilization.LowNodeUtilizationPluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: nodeutilization.LowNodeUtilizationPluginName,
+								Args: &nodeutilization.LowNodeUtilizationArgs{
+									UseDeviationThresholds: true,
+									NumberOfNodes:          3,
+									Thresholds: api.ResourceThresholds{
+										"cpu":    api.Percentage(20),
+										"memory": api.Percentage(20),
+										"pods":   api.Percentage(20),
+									},
+									TargetThresholds: api.ResourceThresholds{
+										"cpu":    api.Percentage(50),
+										"memory": api.Percentage(50),
+										"pods":   api.Percentage(50),
+									},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{nodeutilization.LowNodeUtilizationPluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removeduplicates.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removeduplicates.PluginName,
+								Args: &removeduplicates.RemoveDuplicatesArgs{
+									ExcludeOwnerKinds: []string{"ReplicaSet"},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{removeduplicates.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removefailedpods.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removefailedpods.PluginName,
+								Args: &removefailedpods.RemoveFailedPodsArgs{
+									ExcludeOwnerKinds:       []string{"Job"},
+									MinPodLifetimeSeconds:   utilpointer.Uint(3600),
+									Reasons:                 []string{"NodeAffinity"},
+									IncludingInitContainers: true,
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removefailedpods.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodshavingtoomanyrestarts.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodshavingtoomanyrestarts.PluginName,
+								Args: &removepodshavingtoomanyrestarts.RemovePodsHavingTooManyRestartsArgs{
+									PodRestartThreshold:     100,
+									IncludingInitContainers: true,
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodshavingtoomanyrestarts.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatinginterpodantiaffinity.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatinginterpodantiaffinity.PluginName,
+								Args: &removepodsviolatinginterpodantiaffinity.RemovePodsViolatingInterPodAntiAffinityArgs{},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodsviolatinginterpodantiaffinity.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatingnodeaffinity.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatingnodeaffinity.PluginName,
+								Args: &removepodsviolatingnodeaffinity.RemovePodsViolatingNodeAffinityArgs{
+									NodeAffinityType: []string{"requiredDuringSchedulingIgnoredDuringExecution"},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodsviolatingnodeaffinity.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatingnodetaints.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatingnodetaints.PluginName,
+								Args: &removepodsviolatingnodetaints.RemovePodsViolatingNodeTaintsArgs{
+									ExcludedTaints: []string{"dedicated=special-user", "reserved"},
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Deschedule: api.PluginSet{
+								Enabled: []string{removepodsviolatingnodetaints.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+					{
+						Name: fmt.Sprintf("strategy-%s-profile", removepodsviolatingtopologyspreadconstraint.PluginName),
+						PluginConfigs: []api.PluginConfig{
+							defaultEvictorPluginConfig,
+							{
+								Name: removepodsviolatingtopologyspreadconstraint.PluginName,
+								Args: &removepodsviolatingtopologyspreadconstraint.RemovePodsViolatingTopologySpreadConstraintArgs{
+									IncludeSoftConstraints: true,
+								},
+							},
+						},
+						Plugins: api.Plugins{
+							Balance: api.PluginSet{
+								Enabled: []string{removepodsviolatingtopologyspreadconstraint.PluginName},
+							},
+							Evict: defaultEvictorPluginSet,
+						},
+					},
+				},
+			},
+		},
+		{
+			description: "invalid strategy name",
+			policy: &v1alpha1.DeschedulerPolicy{Strategies: v1alpha1.StrategyList{
+				"InvalidName": v1alpha1.DeschedulerStrategy{
+					Enabled: true,
+					Params:  &v1alpha1.StrategyParameters{},
+				},
+			}},
+			result: nil,
+			err:    fmt.Errorf("unknown strategy name: InvalidName"),
+		},
+		{
+			description: "invalid threshold priority",
+			policy: &v1alpha1.DeschedulerPolicy{Strategies: v1alpha1.StrategyList{
+				nodeutilization.LowNodeUtilizationPluginName: v1alpha1.DeschedulerStrategy{
+					Enabled: true,
+					Params: &v1alpha1.StrategyParameters{
+						ThresholdPriority:          utilpointer.Int32(100),
+						ThresholdPriorityClassName: "name",
+						NodeResourceUtilizationThresholds: &v1alpha1.NodeResourceUtilizationThresholds{
+							Thresholds: v1alpha1.ResourceThresholds{
+								"cpu":    v1alpha1.Percentage(20),
+								"memory": v1alpha1.Percentage(20),
+								"pods":   v1alpha1.Percentage(20),
+							},
+							TargetThresholds: v1alpha1.ResourceThresholds{
+								"cpu":    v1alpha1.Percentage(50),
+								"memory": v1alpha1.Percentage(50),
+								"pods":   v1alpha1.Percentage(50),
+							},
+						},
+					},
+				},
+			}},
+			result: nil,
+			err:    fmt.Errorf("priority threshold misconfigured for plugin LowNodeUtilization"),
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.description, func(t *testing.T) {
+			client := fakeclientset.NewSimpleClientset()
+			result, err := V1alpha1ToInternal(client, tc.policy, pluginbuilder.PluginRegistry)
+			if err != nil {
+				if err.Error() != tc.err.Error() {
+					t.Errorf("unexpected error: %s", err.Error())
+				}
+			}
+			if err == nil {
+				// sort to easily compare deepequality
+				result.Profiles = api.SortProfilesByName(result.Profiles)
+				diff := cmp.Diff(tc.result, result)
+				if diff != "" {
+					t.Errorf("test '%s' failed. Results are not deep equal. mismatch (-want +got):\n%s", tc.description, diff)
+				}
+			}
+		})
+	}
+}

pkg/descheduler/setupplugins.go

@@ -0,0 +1,50 @@
+/*
+Copyright 2017 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package descheduler
+
+import (
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/defaultevictor"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/nodeutilization"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/pluginbuilder"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/podlifetime"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removeduplicates"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removefailedpods"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodshavingtoomanyrestarts"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatinginterpodantiaffinity"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingnodeaffinity"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingnodetaints"
+	"sigs.k8s.io/descheduler/pkg/framework/plugins/removepodsviolatingtopologyspreadconstraint"
+)
+
+func SetupPlugins() {
+	pluginbuilder.PluginRegistry = pluginbuilder.NewRegistry()
+	RegisterDefaultPlugins(pluginbuilder.PluginRegistry)
+}
+
+func RegisterDefaultPlugins(registry pluginbuilder.Registry) {
+	pluginbuilder.Register(defaultevictor.PluginName, defaultevictor.New, &defaultevictor.DefaultEvictorArgs{}, registry)
+	pluginbuilder.Register(nodeutilization.LowNodeUtilizationPluginName, nodeutilization.NewLowNodeUtilization, &nodeutilization.LowNodeUtilizationArgs{}, registry)
+	pluginbuilder.Register(nodeutilization.HighNodeUtilizationPluginName, nodeutilization.NewHighNodeUtilization, &nodeutilization.HighNodeUtilizationArgs{}, registry)
+	pluginbuilder.Register(podlifetime.PluginName, podlifetime.New, &podlifetime.PodLifeTimeArgs{}, registry)
+	pluginbuilder.Register(removeduplicates.PluginName, removeduplicates.New, &removeduplicates.RemoveDuplicatesArgs{}, registry)
+	pluginbuilder.Register(removefailedpods.PluginName, removefailedpods.New, &removefailedpods.RemoveFailedPodsArgs{}, registry)
+	pluginbuilder.Register(removepodshavingtoomanyrestarts.PluginName, removepodshavingtoomanyrestarts.New, &removepodshavingtoomanyrestarts.RemovePodsHavingTooManyRestartsArgs{}, registry)
+	pluginbuilder.Register(removepodsviolatinginterpodantiaffinity.PluginName, removepodsviolatinginterpodantiaffinity.New, &removepodsviolatinginterpodantiaffinity.RemovePodsViolatingInterPodAntiAffinityArgs{}, registry)
+	pluginbuilder.Register(removepodsviolatingnodeaffinity.PluginName, removepodsviolatingnodeaffinity.New, &removepodsviolatingnodeaffinity.RemovePodsViolatingNodeAffinityArgs{}, registry)
+	pluginbuilder.Register(removepodsviolatingnodetaints.PluginName, removepodsviolatingnodetaints.New, &removepodsviolatingnodetaints.RemovePodsViolatingNodeTaintsArgs{}, registry)
+	pluginbuilder.Register(removepodsviolatingtopologyspreadconstraint.PluginName, removepodsviolatingtopologyspreadconstraint.New, &removepodsviolatingtopologyspreadconstraint.RemovePodsViolatingTopologySpreadConstraintArgs{}, registry)
+}

pkg/descheduler/strategies/failedpods.go

@@ -1,163 +0,0 @@
-package strategies
-
-import (
-	"context"
-	"fmt"
-	"k8s.io/apimachinery/pkg/util/sets"
-
-	v1 "k8s.io/api/core/v1"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	utilerrors "k8s.io/apimachinery/pkg/util/errors"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/descheduler/strategies/validation"
-)
-
-// validatedFailedPodsStrategyParams contains validated strategy parameters
-type validatedFailedPodsStrategyParams struct {
-	*validation.ValidatedStrategyParams
-	includingInitContainers bool
-	reasons                 sets.String
-	excludeOwnerKinds       sets.String
-	minPodLifetimeSeconds   *uint
-}
-
-// RemoveFailedPods removes Pods that are in failed status phase.
-func RemoveFailedPods(
-	ctx context.Context,
-	client clientset.Interface,
-	strategy api.DeschedulerStrategy,
-	nodes []*v1.Node,
-	podEvictor *evictions.PodEvictor,
-) {
-	strategyParams, err := validateAndParseRemoveFailedPodsParams(ctx, client, strategy.Params)
-	if err != nil {
-		klog.ErrorS(err, "Invalid RemoveFailedPods parameters")
-		return
-	}
-
-	evictable := podEvictor.Evictable(
-		evictions.WithPriorityThreshold(strategyParams.ThresholdPriority),
-		evictions.WithNodeFit(strategyParams.NodeFit),
-		evictions.WithLabelSelector(strategyParams.LabelSelector),
-	)
-
-	for _, node := range nodes {
-		klog.V(1).InfoS("Processing node", "node", klog.KObj(node))
-		fieldSelectorString := "spec.nodeName=" + node.Name + ",status.phase=" + string(v1.PodFailed)
-		pods, err := podutil.ListPodsOnANodeWithFieldSelector(
-			ctx,
-			client,
-			node,
-			fieldSelectorString,
-			podutil.WithFilter(evictable.IsEvictable),
-			podutil.WithNamespaces(strategyParams.IncludedNamespaces.UnsortedList()),
-			podutil.WithoutNamespaces(strategyParams.ExcludedNamespaces.UnsortedList()),
-			podutil.WithLabelSelector(strategy.Params.LabelSelector),
-		)
-		if err != nil {
-			klog.ErrorS(err, "Error listing a nodes failed pods", "node", klog.KObj(node))
-			continue
-		}
-
-		for i, pod := range pods {
-			if err = validateFailedPodShouldEvict(pod, *strategyParams); err != nil {
-				klog.V(4).InfoS(fmt.Sprintf("ignoring pod for eviction due to: %s", err.Error()), "pod", klog.KObj(pod))
-				continue
-			}
-
-			if _, err = podEvictor.EvictPod(ctx, pods[i], node, "FailedPod"); err != nil {
-				klog.ErrorS(err, "Error evicting pod", "pod", klog.KObj(pod))
-				break
-			}
-		}
-	}
-}
-
-func validateAndParseRemoveFailedPodsParams(
-	ctx context.Context,
-	client clientset.Interface,
-	params *api.StrategyParameters,
-) (*validatedFailedPodsStrategyParams, error) {
-	if params == nil {
-		return &validatedFailedPodsStrategyParams{}, nil
-	}
-
-	strategyParams, err := validation.ValidateAndParseStrategyParams(ctx, client, params)
-	if err != nil {
-		return nil, err
-	}
-
-	var reasons, excludeOwnerKinds sets.String
-	var includingInitContainers bool
-	var minPodLifetimeSeconds *uint
-	if params.FailedPods != nil {
-		reasons = sets.NewString(params.FailedPods.Reasons...)
-		includingInitContainers = params.FailedPods.IncludingInitContainers
-		excludeOwnerKinds = sets.NewString(params.FailedPods.ExcludeOwnerKinds...)
-		minPodLifetimeSeconds = params.FailedPods.MinPodLifetimeSeconds
-	}
-
-	return &validatedFailedPodsStrategyParams{
-		ValidatedStrategyParams: strategyParams,
-		includingInitContainers: includingInitContainers,
-		reasons:                 reasons,
-		excludeOwnerKinds:       excludeOwnerKinds,
-		minPodLifetimeSeconds:   minPodLifetimeSeconds,
-	}, nil
-}
-
-// validateFailedPodShouldEvict looks at strategy params settings to see if the Pod
-// should be evicted given the params in the PodFailed policy.
-func validateFailedPodShouldEvict(pod *v1.Pod, strategyParams validatedFailedPodsStrategyParams) error {
-	var errs []error
-
-	if strategyParams.minPodLifetimeSeconds != nil {
-		podAgeSeconds := uint(metav1.Now().Sub(pod.GetCreationTimestamp().Local()).Seconds())
-		if podAgeSeconds < *strategyParams.minPodLifetimeSeconds {
-			errs = append(errs, fmt.Errorf("pod does not exceed the min age seconds of %d", *strategyParams.minPodLifetimeSeconds))
-		}
-	}
-
-	if len(strategyParams.excludeOwnerKinds) > 0 {
-		ownerRefList := podutil.OwnerRef(pod)
-		for _, owner := range ownerRefList {
-			if strategyParams.excludeOwnerKinds.Has(owner.Kind) {
-				errs = append(errs, fmt.Errorf("pod's owner kind of %s is excluded", owner.Kind))
-			}
-		}
-	}
-
-	if len(strategyParams.reasons) > 0 {
-		reasons := getFailedContainerStatusReasons(pod.Status.ContainerStatuses)
-
-		if strategyParams.includingInitContainers {
-			reasons = append(reasons, getFailedContainerStatusReasons(pod.Status.InitContainerStatuses)...)
-		}
-
-		if !strategyParams.reasons.HasAny(reasons...) {
-			errs = append(errs, fmt.Errorf("pod does not match any of the reasons"))
-		}
-	}
-
-	return utilerrors.NewAggregate(errs)
-}
-
-func getFailedContainerStatusReasons(containerStatuses []v1.ContainerStatus) []string {
-	reasons := make([]string, 0)
-
-	for _, containerStatus := range containerStatuses {
-		if containerStatus.State.Waiting != nil && containerStatus.State.Waiting.Reason != "" {
-			reasons = append(reasons, containerStatus.State.Waiting.Reason)
-		}
-		if containerStatus.State.Terminated != nil && containerStatus.State.Terminated.Reason != "" {
-			reasons = append(reasons, containerStatus.State.Terminated.Reason)
-		}
-	}
-
-	return reasons
-}

pkg/descheduler/strategies/failedpods_test.go

@@ -1,274 +0,0 @@
-package strategies
-
-import (
-	"context"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"testing"
-
-	v1 "k8s.io/api/core/v1"
-	policyv1 "k8s.io/api/policy/v1"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/test"
-)
-
-var (
-	OneHourInSeconds uint = 3600
-)
-
-func TestRemoveFailedPods(t *testing.T) {
-	ctx := context.Background()
-
-	createStrategy := func(enabled, includingInitContainers bool, reasons, excludeKinds []string, minAgeSeconds *uint, nodeFit bool) api.DeschedulerStrategy {
-		return api.DeschedulerStrategy{
-			Enabled: enabled,
-			Params: &api.StrategyParameters{
-				FailedPods: &api.FailedPods{
-					Reasons:                 reasons,
-					IncludingInitContainers: includingInitContainers,
-					ExcludeOwnerKinds:       excludeKinds,
-					MinPodLifetimeSeconds:   minAgeSeconds,
-				},
-				NodeFit: nodeFit,
-			},
-		}
-	}
-
-	tests := []struct {
-		description             string
-		nodes                   []*v1.Node
-		strategy                api.DeschedulerStrategy
-		expectedEvictedPodCount int
-		pods                    []v1.Pod
-	}{
-		{
-			description:             "0 failures, 0 evictions",
-			strategy:                createStrategy(true, false, nil, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 0,
-			pods:                    []v1.Pod{}, // no pods come back with field selector phase=Failed
-		},
-		{
-			description:             "1 container terminated with reason NodeAffinity, 1 eviction",
-			strategy:                createStrategy(true, false, nil, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 1,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", nil, &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}),
-			},
-		},
-		{
-			description:             "1 init container terminated with reason NodeAffinity, 1 eviction",
-			strategy:                createStrategy(true, true, nil, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 1,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}, nil),
-			},
-		},
-		{
-			description:             "1 init container waiting with reason CreateContainerConfigError, 1 eviction",
-			strategy:                createStrategy(true, true, nil, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 1,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", &v1.ContainerState{
-					Waiting: &v1.ContainerStateWaiting{Reason: "CreateContainerConfigError"},
-				}, nil),
-			},
-		},
-		{
-			description: "2 init container waiting with reason CreateContainerConfigError, 2 nodes, 2 evictions",
-			strategy:    createStrategy(true, true, nil, nil, nil, false),
-			nodes: []*v1.Node{
-				test.BuildTestNode("node1", 2000, 3000, 10, nil),
-				test.BuildTestNode("node2", 2000, 3000, 10, nil),
-			},
-			expectedEvictedPodCount: 2,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "CreateContainerConfigError"},
-				}, nil),
-				buildTestPod("p2", "node2", &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "CreateContainerConfigError"},
-				}, nil),
-			},
-		},
-		{
-			description:             "include reason=CreateContainerConfigError, 1 container terminated with reason CreateContainerConfigError, 1 eviction",
-			strategy:                createStrategy(true, false, []string{"CreateContainerConfigError"}, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 1,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", nil, &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "CreateContainerConfigError"},
-				}),
-			},
-		},
-		{
-			description:             "include reason=CreateContainerConfigError+NodeAffinity, 1 container terminated with reason CreateContainerConfigError, 1 eviction",
-			strategy:                createStrategy(true, false, []string{"CreateContainerConfigError", "NodeAffinity"}, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 1,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", nil, &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "CreateContainerConfigError"},
-				}),
-			},
-		},
-		{
-			description:             "include reason=CreateContainerConfigError, 1 container terminated with reason NodeAffinity, 0 eviction",
-			strategy:                createStrategy(true, false, []string{"CreateContainerConfigError"}, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 0,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", nil, &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}),
-			},
-		},
-		{
-			description:             "include init container=false, 1 init container waiting with reason CreateContainerConfigError, 0 eviction",
-			strategy:                createStrategy(true, false, []string{"CreateContainerConfigError"}, nil, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 0,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", &v1.ContainerState{
-					Waiting: &v1.ContainerStateWaiting{Reason: "CreateContainerConfigError"},
-				}, nil),
-			},
-		},
-		{
-			description:             "lifetime 1 hour, 1 container terminated with reason NodeAffinity, 0 eviction",
-			strategy:                createStrategy(true, false, nil, nil, &OneHourInSeconds, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 0,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", nil, &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}),
-			},
-		},
-		{
-			description: "nodeFit=true, 1 unschedulable node, 1 container terminated with reason NodeAffinity, 0 eviction",
-			strategy:    createStrategy(true, false, nil, nil, nil, true),
-			nodes: []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, func(node *v1.Node) {
-				node.Spec.Unschedulable = true
-			})},
-			expectedEvictedPodCount: 0,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", nil, &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}),
-			},
-		},
-		{
-			description:             "excluded owner kind=ReplicaSet, 1 init container terminated with owner kind=ReplicaSet, 0 eviction",
-			strategy:                createStrategy(true, true, nil, []string{"ReplicaSet"}, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 0,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}, nil),
-			},
-		},
-		{
-			description:             "excluded owner kind=DaemonSet, 1 init container terminated with owner kind=ReplicaSet, 1 eviction",
-			strategy:                createStrategy(true, true, nil, []string{"DaemonSet"}, nil, false),
-			nodes:                   []*v1.Node{test.BuildTestNode("node1", 2000, 3000, 10, nil)},
-			expectedEvictedPodCount: 1,
-			pods: []v1.Pod{
-				buildTestPod("p1", "node1", &v1.ContainerState{
-					Terminated: &v1.ContainerStateTerminated{Reason: "NodeAffinity"},
-				}, nil),
-			},
-		},
-	}
-	for _, tc := range tests {
-		fakeClient := &fake.Clientset{}
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			return true, &v1.PodList{Items: tc.pods}, nil
-		})
-
-		podEvictor := evictions.NewPodEvictor(
-			fakeClient,
-			policyv1.SchemeGroupVersion.String(),
-			false,
-			100,
-			tc.nodes,
-			false,
-			false,
-			false,
-		)
-
-		RemoveFailedPods(ctx, fakeClient, tc.strategy, tc.nodes, podEvictor)
-		actualEvictedPodCount := podEvictor.TotalEvicted()
-		if actualEvictedPodCount != tc.expectedEvictedPodCount {
-			t.Errorf("Test %#v failed, expected %v pod evictions, but got %v pod evictions\n", tc.description, tc.expectedEvictedPodCount, actualEvictedPodCount)
-		}
-	}
-}
-
-func TestValidRemoveFailedPodsParams(t *testing.T) {
-	ctx := context.Background()
-	fakeClient := &fake.Clientset{}
-	testCases := []struct {
-		name   string
-		params *api.StrategyParameters
-	}{
-		{name: "validate nil params", params: nil},
-		{name: "validate reasons params", params: &api.StrategyParameters{FailedPods: &api.FailedPods{
-			Reasons: []string{"CreateContainerConfigError"},
-		}}},
-		{name: "validate includingInitContainers params", params: &api.StrategyParameters{FailedPods: &api.FailedPods{
-			IncludingInitContainers: true,
-		}}},
-		{name: "validate excludeOwnerKinds params", params: &api.StrategyParameters{FailedPods: &api.FailedPods{
-			ExcludeOwnerKinds: []string{"Job"},
-		}}},
-		{name: "validate excludeOwnerKinds params", params: &api.StrategyParameters{FailedPods: &api.FailedPods{
-			MinPodLifetimeSeconds: &OneHourInSeconds,
-		}}},
-	}
-	for _, tc := range testCases {
-		t.Run(tc.name, func(t *testing.T) {
-			t.Parallel()
-			params, err := validateAndParseRemoveFailedPodsParams(ctx, fakeClient, tc.params)
-			if err != nil {
-				t.Errorf("strategy params should be valid but got err: %v", err.Error())
-			}
-			if params == nil {
-				t.Errorf("strategy params should return a ValidatedFailedPodsStrategyParams but got nil")
-			}
-		})
-	}
-}
-
-func buildTestPod(podName, nodeName string, initContainerState, containerState *v1.ContainerState) v1.Pod {
-	pod := test.BuildTestPod(podName, 1, 1, nodeName, func(p *v1.Pod) {
-		ps := v1.PodStatus{}
-
-		if initContainerState != nil {
-			ps.InitContainerStatuses = []v1.ContainerStatus{{State: *initContainerState}}
-			ps.Phase = v1.PodFailed
-		}
-
-		if containerState != nil {
-			ps.ContainerStatuses = []v1.ContainerStatus{{State: *containerState}}
-			ps.Phase = v1.PodFailed
-		}
-
-		p.Status = ps
-	})
-	pod.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	pod.ObjectMeta.SetCreationTimestamp(metav1.Now())
-	return *pod
-}

pkg/descheduler/strategies/node_affinity.go

@@ -1,113 +0,0 @@
-/*
-Copyright 2017 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"fmt"
-
-	v1 "k8s.io/api/core/v1"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	nodeutil "sigs.k8s.io/descheduler/pkg/descheduler/node"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/utils"
-)
-
-func validatePodsViolatingNodeAffinityParams(params *api.StrategyParameters) error {
-	if params == nil || len(params.NodeAffinityType) == 0 {
-		return fmt.Errorf("NodeAffinityType is empty")
-	}
-	// At most one of include/exclude can be set
-	if params.Namespaces != nil && len(params.Namespaces.Include) > 0 && len(params.Namespaces.Exclude) > 0 {
-		return fmt.Errorf("only one of Include/Exclude namespaces can be set")
-	}
-	if params.ThresholdPriority != nil && params.ThresholdPriorityClassName != "" {
-		return fmt.Errorf("only one of thresholdPriority and thresholdPriorityClassName can be set")
-	}
-
-	return nil
-}
-
-// RemovePodsViolatingNodeAffinity evicts pods on nodes which violate node affinity
-func RemovePodsViolatingNodeAffinity(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, podEvictor *evictions.PodEvictor) {
-	if err := validatePodsViolatingNodeAffinityParams(strategy.Params); err != nil {
-		klog.ErrorS(err, "Invalid RemovePodsViolatingNodeAffinity parameters")
-		return
-	}
-	thresholdPriority, err := utils.GetPriorityFromStrategyParams(ctx, client, strategy.Params)
-	if err != nil {
-		klog.ErrorS(err, "Failed to get threshold priority from strategy's params")
-		return
-	}
-
-	var includedNamespaces, excludedNamespaces []string
-	if strategy.Params.Namespaces != nil {
-		includedNamespaces = strategy.Params.Namespaces.Include
-		excludedNamespaces = strategy.Params.Namespaces.Exclude
-	}
-
-	nodeFit := false
-	if strategy.Params != nil {
-		nodeFit = strategy.Params.NodeFit
-	}
-
-	evictable := podEvictor.Evictable(evictions.WithPriorityThreshold(thresholdPriority), evictions.WithNodeFit(nodeFit))
-
-	for _, nodeAffinity := range strategy.Params.NodeAffinityType {
-		klog.V(2).InfoS("Executing for nodeAffinityType", "nodeAffinity", nodeAffinity)
-
-		switch nodeAffinity {
-		case "requiredDuringSchedulingIgnoredDuringExecution":
-			for _, node := range nodes {
-				klog.V(1).InfoS("Processing node", "node", klog.KObj(node))
-
-				pods, err := podutil.ListPodsOnANode(
-					ctx,
-					client,
-					node,
-					podutil.WithFilter(func(pod *v1.Pod) bool {
-						return evictable.IsEvictable(pod) &&
-							!nodeutil.PodFitsCurrentNode(pod, node) &&
-							nodeutil.PodFitsAnyNode(pod, nodes)
-					}),
-					podutil.WithNamespaces(includedNamespaces),
-					podutil.WithoutNamespaces(excludedNamespaces),
-					podutil.WithLabelSelector(strategy.Params.LabelSelector),
-				)
-				if err != nil {
-					klog.ErrorS(err, "Failed to get pods", "node", klog.KObj(node))
-				}
-
-				for _, pod := range pods {
-					if pod.Spec.Affinity != nil && pod.Spec.Affinity.NodeAffinity != nil && pod.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution != nil {
-						klog.V(1).InfoS("Evicting pod", "pod", klog.KObj(pod))
-						if _, err := podEvictor.EvictPod(ctx, pod, node, "NodeAffinity"); err != nil {
-							klog.ErrorS(err, "Error evicting pod")
-							break
-						}
-					}
-				}
-			}
-		default:
-			klog.ErrorS(nil, "Invalid nodeAffinityType", "nodeAffinity", nodeAffinity)
-		}
-	}
-}

pkg/descheduler/strategies/node_affinity_test.go

@@ -1,190 +0,0 @@
-/*
-Copyright 2017 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"testing"
-
-	v1 "k8s.io/api/core/v1"
-	policyv1 "k8s.io/api/policy/v1"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/test"
-)
-
-func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
-	ctx := context.Background()
-	requiredDuringSchedulingIgnoredDuringExecutionStrategy := api.DeschedulerStrategy{
-		Enabled: true,
-		Params: &api.StrategyParameters{
-			NodeAffinityType: []string{
-				"requiredDuringSchedulingIgnoredDuringExecution",
-			},
-		},
-	}
-
-	requiredDuringSchedulingIgnoredDuringExecutionWithNodeFitStrategy := api.DeschedulerStrategy{
-		Enabled: true,
-		Params: &api.StrategyParameters{
-			NodeAffinityType: []string{
-				"requiredDuringSchedulingIgnoredDuringExecution",
-			},
-			NodeFit: true,
-		},
-	}
-
-	nodeLabelKey := "kubernetes.io/desiredNode"
-	nodeLabelValue := "yes"
-	nodeWithLabels := test.BuildTestNode("nodeWithLabels", 2000, 3000, 10, nil)
-	nodeWithLabels.Labels[nodeLabelKey] = nodeLabelValue
-
-	nodeWithoutLabels := test.BuildTestNode("nodeWithoutLabels", 2000, 3000, 10, nil)
-
-	unschedulableNodeWithLabels := test.BuildTestNode("unschedulableNodeWithLabels", 2000, 3000, 10, nil)
-	nodeWithLabels.Labels[nodeLabelKey] = nodeLabelValue
-	unschedulableNodeWithLabels.Spec.Unschedulable = true
-
-	addPodsToNode := func(node *v1.Node) []v1.Pod {
-		podWithNodeAffinity := test.BuildTestPod("podWithNodeAffinity", 100, 0, node.Name, nil)
-		podWithNodeAffinity.Spec.Affinity = &v1.Affinity{
-			NodeAffinity: &v1.NodeAffinity{
-				RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
-					NodeSelectorTerms: []v1.NodeSelectorTerm{
-						{
-							MatchExpressions: []v1.NodeSelectorRequirement{
-								{
-									Key:      nodeLabelKey,
-									Operator: "In",
-									Values: []string{
-										nodeLabelValue,
-									},
-								},
-							},
-						},
-					},
-				},
-			},
-		}
-
-		pod1 := test.BuildTestPod("pod1", 100, 0, node.Name, nil)
-		pod2 := test.BuildTestPod("pod2", 100, 0, node.Name, nil)
-
-		podWithNodeAffinity.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-		pod1.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-		pod2.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-		return []v1.Pod{
-			*podWithNodeAffinity,
-			*pod1,
-			*pod2,
-		}
-	}
-
-	tests := []struct {
-		description             string
-		nodes                   []*v1.Node
-		pods                    []v1.Pod
-		strategy                api.DeschedulerStrategy
-		expectedEvictedPodCount int
-		maxPodsToEvictPerNode   int
-	}{
-		{
-			description: "Invalid strategy type, should not evict any pods",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					NodeAffinityType: []string{
-						"requiredDuringSchedulingRequiredDuringExecution",
-					},
-				},
-			},
-			expectedEvictedPodCount: 0,
-			pods:                    addPodsToNode(nodeWithoutLabels),
-			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Pod is correctly scheduled on node, no eviction expected",
-			strategy:                requiredDuringSchedulingIgnoredDuringExecutionStrategy,
-			expectedEvictedPodCount: 0,
-			pods:                    addPodsToNode(nodeWithLabels),
-			nodes:                   []*v1.Node{nodeWithLabels},
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Pod is scheduled on node without matching labels, another schedulable node available, should be evicted",
-			expectedEvictedPodCount: 1,
-			strategy:                requiredDuringSchedulingIgnoredDuringExecutionStrategy,
-			pods:                    addPodsToNode(nodeWithoutLabels),
-			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Pod is scheduled on node without matching labels, another schedulable node available, maxPodsToEvictPerNode set to 1, should not be evicted",
-			expectedEvictedPodCount: 1,
-			strategy:                requiredDuringSchedulingIgnoredDuringExecutionStrategy,
-			pods:                    addPodsToNode(nodeWithoutLabels),
-			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			maxPodsToEvictPerNode:   1,
-		},
-		{
-			description:             "Pod is scheduled on node without matching labels, but no node where pod fits is available, should not evict",
-			expectedEvictedPodCount: 0,
-			strategy:                requiredDuringSchedulingIgnoredDuringExecutionWithNodeFitStrategy,
-			pods:                    addPodsToNode(nodeWithoutLabels),
-			nodes:                   []*v1.Node{nodeWithoutLabels, unschedulableNodeWithLabels},
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Pod is scheduled on node without matching labels, and node where pod fits is available, should evict",
-			expectedEvictedPodCount: 0,
-			strategy:                requiredDuringSchedulingIgnoredDuringExecutionWithNodeFitStrategy,
-			pods:                    addPodsToNode(nodeWithoutLabels),
-			nodes:                   []*v1.Node{nodeWithLabels, unschedulableNodeWithLabels},
-			maxPodsToEvictPerNode:   1,
-		},
-	}
-
-	for _, tc := range tests {
-
-		fakeClient := &fake.Clientset{}
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			return true, &v1.PodList{Items: tc.pods}, nil
-		})
-
-		podEvictor := evictions.NewPodEvictor(
-			fakeClient,
-			policyv1.SchemeGroupVersion.String(),
-			false,
-			tc.maxPodsToEvictPerNode,
-			tc.nodes,
-			false,
-			false,
-			false,
-		)
-
-		RemovePodsViolatingNodeAffinity(ctx, fakeClient, tc.strategy, tc.nodes, podEvictor)
-		actualEvictedPodCount := podEvictor.TotalEvicted()
-		if actualEvictedPodCount != tc.expectedEvictedPodCount {
-			t.Errorf("Test %#v failed, expected %v pod evictions, but got %v pod evictions\n", tc.description, tc.expectedEvictedPodCount, actualEvictedPodCount)
-		}
-	}
-}

pkg/descheduler/strategies/node_taint.go

@@ -1,110 +0,0 @@
-/*
-Copyright 2017 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"fmt"
-
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/utils"
-
-	v1 "k8s.io/api/core/v1"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-)
-
-func validateRemovePodsViolatingNodeTaintsParams(params *api.StrategyParameters) error {
-	if params == nil {
-		return nil
-	}
-
-	// At most one of include/exclude can be set
-	if params.Namespaces != nil && len(params.Namespaces.Include) > 0 && len(params.Namespaces.Exclude) > 0 {
-		return fmt.Errorf("only one of Include/Exclude namespaces can be set")
-	}
-	if params.ThresholdPriority != nil && params.ThresholdPriorityClassName != "" {
-		return fmt.Errorf("only one of thresholdPriority and thresholdPriorityClassName can be set")
-	}
-
-	return nil
-}
-
-// RemovePodsViolatingNodeTaints evicts pods on the node which violate NoSchedule Taints on nodes
-func RemovePodsViolatingNodeTaints(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, podEvictor *evictions.PodEvictor) {
-	if err := validateRemovePodsViolatingNodeTaintsParams(strategy.Params); err != nil {
-		klog.ErrorS(err, "Invalid RemovePodsViolatingNodeTaints parameters")
-		return
-	}
-
-	var includedNamespaces, excludedNamespaces []string
-	var labelSelector *metav1.LabelSelector
-	if strategy.Params != nil {
-		if strategy.Params.Namespaces != nil {
-			includedNamespaces = strategy.Params.Namespaces.Include
-			excludedNamespaces = strategy.Params.Namespaces.Exclude
-		}
-		labelSelector = strategy.Params.LabelSelector
-	}
-
-	thresholdPriority, err := utils.GetPriorityFromStrategyParams(ctx, client, strategy.Params)
-	if err != nil {
-		klog.ErrorS(err, "Failed to get threshold priority from strategy's params")
-		return
-	}
-
-	nodeFit := false
-	if strategy.Params != nil {
-		nodeFit = strategy.Params.NodeFit
-	}
-
-	evictable := podEvictor.Evictable(evictions.WithPriorityThreshold(thresholdPriority), evictions.WithNodeFit(nodeFit))
-
-	for _, node := range nodes {
-		klog.V(1).InfoS("Processing node", "node", klog.KObj(node))
-		pods, err := podutil.ListPodsOnANode(
-			ctx,
-			client,
-			node,
-			podutil.WithFilter(evictable.IsEvictable),
-			podutil.WithNamespaces(includedNamespaces),
-			podutil.WithoutNamespaces(excludedNamespaces),
-			podutil.WithLabelSelector(labelSelector),
-		)
-		if err != nil {
-			//no pods evicted as error encountered retrieving evictable Pods
-			return
-		}
-		totalPods := len(pods)
-		for i := 0; i < totalPods; i++ {
-			if !utils.TolerationsTolerateTaintsWithFilter(
-				pods[i].Spec.Tolerations,
-				node.Spec.Taints,
-				func(taint *v1.Taint) bool { return taint.Effect == v1.TaintEffectNoSchedule },
-			) {
-				klog.V(2).InfoS("Not all taints with NoSchedule effect are tolerated after update for pod on node", "pod", klog.KObj(pods[i]), "node", klog.KObj(node))
-				if _, err := podEvictor.EvictPod(ctx, pods[i], node, "NodeTaint"); err != nil {
-					klog.ErrorS(err, "Error evicting pod")
-					break
-				}
-			}
-		}
-	}
-}

pkg/descheduler/strategies/node_taint_test.go

@@ -1,359 +0,0 @@
-package strategies
-
-import (
-	"context"
-	"fmt"
-	"testing"
-
-	v1 "k8s.io/api/core/v1"
-	policyv1 "k8s.io/api/policy/v1"
-	"k8s.io/apimachinery/pkg/api/resource"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/pkg/utils"
-	"sigs.k8s.io/descheduler/test"
-)
-
-func createNoScheduleTaint(key, value string, index int) v1.Taint {
-	return v1.Taint{
-		Key:    "testTaint" + fmt.Sprintf("%v", index),
-		Value:  "test" + fmt.Sprintf("%v", index),
-		Effect: v1.TaintEffectNoSchedule,
-	}
-}
-
-func addTaintsToNode(node *v1.Node, key, value string, indices []int) *v1.Node {
-	taints := []v1.Taint{}
-	for _, index := range indices {
-		taints = append(taints, createNoScheduleTaint(key, value, index))
-	}
-	node.Spec.Taints = taints
-	return node
-}
-
-func addTolerationToPod(pod *v1.Pod, key, value string, index int) *v1.Pod {
-	if pod.Annotations == nil {
-		pod.Annotations = map[string]string{}
-	}
-
-	pod.Spec.Tolerations = []v1.Toleration{{Key: key + fmt.Sprintf("%v", index), Value: value + fmt.Sprintf("%v", index), Effect: v1.TaintEffectNoSchedule}}
-
-	return pod
-}
-
-func TestDeletePodsViolatingNodeTaints(t *testing.T) {
-	ctx := context.Background()
-	node1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
-	node1 = addTaintsToNode(node1, "testTaint", "test", []int{1})
-	node2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
-	node1 = addTaintsToNode(node2, "testingTaint", "testing", []int{1})
-
-	node3 := test.BuildTestNode("n3", 2000, 3000, 10, func(node *v1.Node) {
-		node.ObjectMeta.Labels = map[string]string{
-			"datacenter": "east",
-		}
-	})
-	node4 := test.BuildTestNode("n4", 2000, 3000, 10, func(node *v1.Node) {
-		node.Spec = v1.NodeSpec{
-			Unschedulable: true,
-		}
-	})
-
-	p1 := test.BuildTestPod("p1", 100, 0, node1.Name, nil)
-	p2 := test.BuildTestPod("p2", 100, 0, node1.Name, nil)
-	p3 := test.BuildTestPod("p3", 100, 0, node1.Name, nil)
-	p4 := test.BuildTestPod("p4", 100, 0, node1.Name, nil)
-	p5 := test.BuildTestPod("p5", 100, 0, node1.Name, nil)
-	p6 := test.BuildTestPod("p6", 100, 0, node1.Name, nil)
-
-	p1.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p2.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p3.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p4.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p5.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p6.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p7 := test.BuildTestPod("p7", 100, 0, node2.Name, nil)
-	p8 := test.BuildTestPod("p8", 100, 0, node2.Name, nil)
-	p9 := test.BuildTestPod("p9", 100, 0, node2.Name, nil)
-	p10 := test.BuildTestPod("p10", 100, 0, node2.Name, nil)
-	p11 := test.BuildTestPod("p11", 100, 0, node2.Name, nil)
-	p11.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p12 := test.BuildTestPod("p11", 100, 0, node2.Name, nil)
-	p12.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-	// The following 4 pods won't get evicted.
-	// A Critical Pod.
-	p7.Namespace = "kube-system"
-	priority := utils.SystemCriticalPriority
-	p7.Spec.Priority = &priority
-	p7.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-	// A daemonset.
-	p8.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
-	// A pod with local storage.
-	p9.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p9.Spec.Volumes = []v1.Volume{
-		{
-			Name: "sample",
-			VolumeSource: v1.VolumeSource{
-				HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-				EmptyDir: &v1.EmptyDirVolumeSource{
-					SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
-			},
-		},
-	}
-	// A Mirror Pod.
-	p10.Annotations = test.GetMirrorPodAnnotation()
-
-	p1 = addTolerationToPod(p1, "testTaint", "test", 1)
-	p3 = addTolerationToPod(p3, "testTaint", "test", 1)
-	p4 = addTolerationToPod(p4, "testTaintX", "testX", 1)
-
-	p12.Spec.NodeSelector = map[string]string{
-		"datacenter": "west",
-	}
-
-	tests := []struct {
-		description             string
-		nodes                   []*v1.Node
-		pods                    []v1.Pod
-		evictLocalStoragePods   bool
-		evictSystemCriticalPods bool
-		maxPodsToEvictPerNode   int
-		expectedEvictedPodCount int
-		nodeFit                 bool
-	}{
-
-		{
-			description:             "Pods not tolerating node taint should be evicted",
-			pods:                    []v1.Pod{*p1, *p2, *p3},
-			nodes:                   []*v1.Node{node1},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 1, //p2 gets evicted
-		},
-		{
-			description:             "Pods with tolerations but not tolerating node taint should be evicted",
-			pods:                    []v1.Pod{*p1, *p3, *p4},
-			nodes:                   []*v1.Node{node1},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 1, //p4 gets evicted
-		},
-		{
-			description:             "Only <maxPodsToEvictPerNode> number of Pods not tolerating node taint should be evicted",
-			pods:                    []v1.Pod{*p1, *p5, *p6},
-			nodes:                   []*v1.Node{node1},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   1,
-			expectedEvictedPodCount: 1, //p5 or p6 gets evicted
-		},
-		{
-			description:             "Critical pods not tolerating node taint should not be evicted",
-			pods:                    []v1.Pod{*p7, *p8, *p9, *p10},
-			nodes:                   []*v1.Node{node2},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 0, //nothing is evicted
-		},
-		{
-			description:             "Critical pods except storage pods not tolerating node taint should not be evicted",
-			pods:                    []v1.Pod{*p7, *p8, *p9, *p10},
-			nodes:                   []*v1.Node{node2},
-			evictLocalStoragePods:   true,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 1, //p9 gets evicted
-		},
-		{
-			description:             "Critical and non critical pods, only non critical pods not tolerating node taint should be evicted",
-			pods:                    []v1.Pod{*p7, *p8, *p10, *p11},
-			nodes:                   []*v1.Node{node2},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 1, //p11 gets evicted
-		},
-		{
-			description:             "Critical and non critical pods, pods not tolerating node taint should be evicted even if they are critical",
-			pods:                    []v1.Pod{*p2, *p7, *p9, *p10},
-			nodes:                   []*v1.Node{node2},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: true,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 2, //p2 and p7 are evicted
-		},
-		{
-			description:             "Pod p2 doesn't tolerate taint on it's node, but also doesn't tolerate taints on other nodes",
-			pods:                    []v1.Pod{*p1, *p2, *p3},
-			nodes:                   []*v1.Node{node1, node2},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 0, //p2 gets evicted
-			nodeFit:                 true,
-		},
-		{
-			description:             "Pod p12 doesn't tolerate taint on it's node, but other nodes don't match it's selector",
-			pods:                    []v1.Pod{*p1, *p3, *p12},
-			nodes:                   []*v1.Node{node1, node3},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 0, //p2 gets evicted
-			nodeFit:                 true,
-		},
-		{
-			description:             "Pod p2 doesn't tolerate taint on it's node, but other nodes are unschedulable",
-			pods:                    []v1.Pod{*p1, *p2, *p3},
-			nodes:                   []*v1.Node{node1, node4},
-			evictLocalStoragePods:   false,
-			evictSystemCriticalPods: false,
-			maxPodsToEvictPerNode:   0,
-			expectedEvictedPodCount: 0, //p2 gets evicted
-			nodeFit:                 true,
-		},
-	}
-
-	for _, tc := range tests {
-
-		// create fake client
-		fakeClient := &fake.Clientset{}
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			return true, &v1.PodList{Items: tc.pods}, nil
-		})
-
-		podEvictor := evictions.NewPodEvictor(
-			fakeClient,
-			policyv1.SchemeGroupVersion.String(),
-			false,
-			tc.maxPodsToEvictPerNode,
-			tc.nodes,
-			tc.evictLocalStoragePods,
-			tc.evictSystemCriticalPods,
-			false,
-		)
-
-		strategy := api.DeschedulerStrategy{
-			Params: &api.StrategyParameters{
-				NodeFit: tc.nodeFit,
-			},
-		}
-
-		RemovePodsViolatingNodeTaints(ctx, fakeClient, strategy, tc.nodes, podEvictor)
-		actualEvictedPodCount := podEvictor.TotalEvicted()
-		if actualEvictedPodCount != tc.expectedEvictedPodCount {
-			t.Errorf("Test %#v failed, Unexpected no of pods evicted: pods evicted: %d, expected: %d", tc.description, actualEvictedPodCount, tc.expectedEvictedPodCount)
-		}
-	}
-
-}
-
-func TestToleratesTaint(t *testing.T) {
-
-	testCases := []struct {
-		description     string
-		toleration      v1.Toleration
-		taint           v1.Taint
-		expectTolerated bool
-	}{
-		{
-			description: "toleration and taint have the same key and effect, and operator is Exists, and taint has no value, expect tolerated",
-			toleration: v1.Toleration{
-				Key:      "foo",
-				Operator: v1.TolerationOpExists,
-				Effect:   v1.TaintEffectNoSchedule,
-			},
-			taint: v1.Taint{
-				Key:    "foo",
-				Effect: v1.TaintEffectNoSchedule,
-			},
-			expectTolerated: true,
-		},
-		{
-			description: "toleration and taint have the same key and effect, and operator is Exists, and taint has some value, expect tolerated",
-			toleration: v1.Toleration{
-				Key:      "foo",
-				Operator: v1.TolerationOpExists,
-				Effect:   v1.TaintEffectNoSchedule,
-			},
-			taint: v1.Taint{
-				Key:    "foo",
-				Value:  "bar",
-				Effect: v1.TaintEffectNoSchedule,
-			},
-			expectTolerated: true,
-		},
-		{
-			description: "toleration and taint have the same effect, toleration has empty key and operator is Exists, means match all taints, expect tolerated",
-			toleration: v1.Toleration{
-				Key:      "",
-				Operator: v1.TolerationOpExists,
-				Effect:   v1.TaintEffectNoSchedule,
-			},
-			taint: v1.Taint{
-				Key:    "foo",
-				Value:  "bar",
-				Effect: v1.TaintEffectNoSchedule,
-			},
-			expectTolerated: true,
-		},
-		{
-			description: "toleration and taint have the same key, effect and value, and operator is Equal, expect tolerated",
-			toleration: v1.Toleration{
-				Key:      "foo",
-				Operator: v1.TolerationOpEqual,
-				Value:    "bar",
-				Effect:   v1.TaintEffectNoSchedule,
-			},
-			taint: v1.Taint{
-				Key:    "foo",
-				Value:  "bar",
-				Effect: v1.TaintEffectNoSchedule,
-			},
-			expectTolerated: true,
-		},
-		{
-			description: "toleration and taint have the same key and effect, but different values, and operator is Equal, expect not tolerated",
-			toleration: v1.Toleration{
-				Key:      "foo",
-				Operator: v1.TolerationOpEqual,
-				Value:    "value1",
-				Effect:   v1.TaintEffectNoSchedule,
-			},
-			taint: v1.Taint{
-				Key:    "foo",
-				Value:  "value2",
-				Effect: v1.TaintEffectNoSchedule,
-			},
-			expectTolerated: false,
-		},
-		{
-			description: "toleration and taint have the same key and value, but different effects, and operator is Equal, expect not tolerated",
-			toleration: v1.Toleration{
-				Key:      "foo",
-				Operator: v1.TolerationOpEqual,
-				Value:    "bar",
-				Effect:   v1.TaintEffectNoSchedule,
-			},
-			taint: v1.Taint{
-				Key:    "foo",
-				Value:  "bar",
-				Effect: v1.TaintEffectNoExecute,
-			},
-			expectTolerated: false,
-		},
-	}
-	for _, tc := range testCases {
-		if tolerated := tc.toleration.ToleratesTaint(&tc.taint); tc.expectTolerated != tolerated {
-			t.Errorf("[%s] expect %v, got %v: toleration %+v, taint %s", tc.description, tc.expectTolerated, tolerated, tc.toleration, tc.taint.ToString())
-		}
-	}
-}

pkg/descheduler/strategies/nodeutilization/highnodeutilization_test.go

@@ -1,784 +0,0 @@
-/*
-Copyright 2021 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-   http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package nodeutilization
-
-import (
-	"context"
-	"fmt"
-	"strings"
-	"testing"
-
-	v1 "k8s.io/api/core/v1"
-	"k8s.io/api/policy/v1beta1"
-	"k8s.io/apimachinery/pkg/api/resource"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/pkg/utils"
-	"sigs.k8s.io/descheduler/test"
-)
-
-func TestHighNodeUtilization(t *testing.T) {
-	ctx := context.Background()
-	n1NodeName := "n1"
-	n2NodeName := "n2"
-	n3NodeName := "n3"
-
-	nodeSelectorKey := "datacenter"
-	nodeSelectorValue := "west"
-
-	testCases := []struct {
-		name                  string
-		thresholds            api.ResourceThresholds
-		nodes                 map[string]*v1.Node
-		pods                  map[string]*v1.PodList
-		maxPodsToEvictPerNode int
-		expectedPodsEvicted   int
-		evictedPods           []string
-	}{
-		{
-			name: "no node below threshold usage",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  20,
-				v1.ResourcePods: 20,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 10, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 4000, 3000, 10, nil),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p3", 400, 0, n1NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p4", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p6", 400, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p7", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p8", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p9", 400, 0, n3NodeName, test.SetRSOwnerRef),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   0,
-		},
-		{
-			name: "no evictable pods",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  40,
-				v1.ResourcePods: 40,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 9, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 4000, 3000, 10, nil),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							// A pod with local storage.
-							test.SetNormalOwnerRef(pod)
-							pod.Spec.Volumes = []v1.Volume{
-								{
-									Name: "sample",
-									VolumeSource: v1.VolumeSource{
-										HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-										EmptyDir: &v1.EmptyDirVolumeSource{
-											SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
-									},
-								},
-							}
-							// A Mirror Pod.
-							pod.Annotations = test.GetMirrorPodAnnotation()
-						}),
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							// A Critical Pod.
-							pod.Namespace = "kube-system"
-							priority := utils.SystemCriticalPriority
-							pod.Spec.Priority = &priority
-						}),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p3", 400, 0, n2NodeName, test.SetDSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n2NodeName, test.SetDSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p5", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p6", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p7", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p8", 400, 0, n3NodeName, test.SetRSOwnerRef),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   0,
-		},
-		{
-			name: "no node to schedule evicted pods",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  20,
-				v1.ResourcePods: 20,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 10, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 4000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						// These can't be evicted.
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These can't be evicted.
-						*test.BuildTestPod("p2", 400, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p3", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n3NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p5", 400, 0, n3NodeName, test.SetRSOwnerRef),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   0,
-		},
-		{
-			name: "without priorities",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  30,
-				v1.ResourcePods: 30,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 10, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 4000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						// These won't be evicted.
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							// A Critical Pod.
-							pod.Namespace = "kube-system"
-							priority := utils.SystemCriticalPriority
-							pod.Spec.Priority = &priority
-						}),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p3", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p6", 400, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p7", 400, 0, n3NodeName, test.SetRSOwnerRef),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   2,
-			evictedPods:           []string{"p1", "p7"},
-		},
-		{
-			name: "without priorities stop when resource capacity is depleted",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  30,
-				v1.ResourcePods: 30,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 2000, 3000, 10, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 2000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 2000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p2", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p3", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p6", 400, 0, n3NodeName, test.SetRSOwnerRef),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   1,
-		},
-		{
-			name: "with priorities",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU: 30,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 10, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 2000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 2000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodPriority(pod, lowPriority)
-						}),
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodPriority(pod, highPriority)
-						}),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p6", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p7", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p8", 400, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p9", 400, 0, n3NodeName, test.SetDSOwnerRef),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   1,
-			evictedPods:           []string{"p1"},
-		},
-		{
-			name: "without priorities evicting best-effort pods only",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU: 30,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 3000, 3000, 10, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 3000, 3000, 5, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 3000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			// All pods are assumed to be burstable (test.BuildTestNode always sets both cpu/memory resource requests to some value)
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.MakeBestEffortPod(pod)
-						}),
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-						}),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						// These won't be evicted.
-						*test.BuildTestPod("p3", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p6", 400, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   1,
-			evictedPods:           []string{"p1"},
-		},
-		{
-			name: "with extended resource",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:   20,
-				extendedResource: 40,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 10, func(node *v1.Node) {
-					test.SetNodeExtendedResource(node, extendedResource, 8)
-				}),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, func(node *v1.Node) {
-					test.SetNodeExtendedResource(node, extendedResource, 8)
-				}),
-				n3NodeName: test.BuildTestNode(n3NodeName, 4000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 100, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-						*test.BuildTestPod("p2", 100, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-						// These won't be evicted
-						*test.BuildTestPod("p2", 100, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetDSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p3", 500, 0, n2NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-						*test.BuildTestPod("p4", 500, 0, n2NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-						*test.BuildTestPod("p5", 500, 0, n2NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-						*test.BuildTestPod("p6", 500, 0, n2NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   2,
-			evictedPods:           []string{"p1", "p2"},
-		},
-		{
-			name: "with extended resource in some of nodes",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:   40,
-				extendedResource: 40,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 10, func(node *v1.Node) {
-					test.SetNodeExtendedResource(node, extendedResource, 8)
-				}),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-				n3NodeName: test.BuildTestNode(n3NodeName, 4000, 3000, 10, test.SetNodeUnschedulable),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						//These won't be evicted
-						*test.BuildTestPod("p1", 100, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-						*test.BuildTestPod("p2", 100, 0, n1NodeName, func(pod *v1.Pod) {
-							test.SetRSOwnerRef(pod)
-							test.SetPodExtendedResourceRequest(pod, extendedResource, 1)
-						}),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p3", 500, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 500, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p5", 500, 0, n2NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p6", 500, 0, n2NodeName, test.SetRSOwnerRef),
-					},
-				},
-				n3NodeName: {
-					Items: []v1.Pod{},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   0,
-		},
-		{
-			name: "Other node match pod node selector",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  30,
-				v1.ResourcePods: 30,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 9, func(node *v1.Node) {
-					node.ObjectMeta.Labels = map[string]string{
-						nodeSelectorKey: nodeSelectorValue,
-					}
-				}),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p3", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n1NodeName, test.SetDSOwnerRef),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, func(pod *v1.Pod) {
-							// A pod selecting nodes in the "west" datacenter
-							test.SetRSOwnerRef(pod)
-							pod.Spec.NodeSelector = map[string]string{
-								nodeSelectorKey: nodeSelectorValue,
-							}
-						}),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   1,
-		},
-		{
-			name: "Other node does not match pod node selector",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:  30,
-				v1.ResourcePods: 30,
-			},
-			nodes: map[string]*v1.Node{
-				n1NodeName: test.BuildTestNode(n1NodeName, 4000, 3000, 9, nil),
-				n2NodeName: test.BuildTestNode(n2NodeName, 4000, 3000, 10, nil),
-			},
-			pods: map[string]*v1.PodList{
-				n1NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p1", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p2", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p3", 400, 0, n1NodeName, test.SetRSOwnerRef),
-						*test.BuildTestPod("p4", 400, 0, n1NodeName, test.SetDSOwnerRef),
-					},
-				},
-				n2NodeName: {
-					Items: []v1.Pod{
-						*test.BuildTestPod("p5", 400, 0, n2NodeName, func(pod *v1.Pod) {
-							// A pod selecting nodes in the "west" datacenter
-							test.SetRSOwnerRef(pod)
-							pod.Spec.NodeSelector = map[string]string{
-								nodeSelectorKey: nodeSelectorValue,
-							}
-						}),
-					},
-				},
-			},
-			maxPodsToEvictPerNode: 0,
-			expectedPodsEvicted:   0,
-		},
-	}
-
-	for _, test := range testCases {
-		t.Run(test.name, func(t *testing.T) {
-			fakeClient := &fake.Clientset{}
-			fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-				list := action.(core.ListAction)
-				fieldString := list.GetListRestrictions().Fields.String()
-				if strings.Contains(fieldString, n1NodeName) {
-					return true, test.pods[n1NodeName], nil
-				}
-				if strings.Contains(fieldString, n2NodeName) {
-					return true, test.pods[n2NodeName], nil
-				}
-				if strings.Contains(fieldString, n3NodeName) {
-					return true, test.pods[n3NodeName], nil
-				}
-				return true, nil, fmt.Errorf("Failed to list: %v", list)
-			})
-			fakeClient.Fake.AddReactor("get", "nodes", func(action core.Action) (bool, runtime.Object, error) {
-				getAction := action.(core.GetAction)
-				if node, exists := test.nodes[getAction.GetName()]; exists {
-					return true, node, nil
-				}
-				return true, nil, fmt.Errorf("Wrong node: %v", getAction.GetName())
-			})
-			podsForEviction := make(map[string]struct{})
-			for _, pod := range test.evictedPods {
-				podsForEviction[pod] = struct{}{}
-			}
-
-			evictionFailed := false
-			if len(test.evictedPods) > 0 {
-				fakeClient.Fake.AddReactor("create", "pods", func(action core.Action) (bool, runtime.Object, error) {
-					getAction := action.(core.CreateAction)
-					obj := getAction.GetObject()
-					if eviction, ok := obj.(*v1beta1.Eviction); ok {
-						if _, exists := podsForEviction[eviction.Name]; exists {
-							return true, obj, nil
-						}
-						evictionFailed = true
-						return true, nil, fmt.Errorf("pod %q was unexpectedly evicted", eviction.Name)
-					}
-					return true, obj, nil
-				})
-			}
-
-			var nodes []*v1.Node
-			for _, node := range test.nodes {
-				nodes = append(nodes, node)
-			}
-
-			podEvictor := evictions.NewPodEvictor(
-				fakeClient,
-				"v1",
-				false,
-				test.maxPodsToEvictPerNode,
-				nodes,
-				false,
-				false,
-				false,
-			)
-
-			strategy := api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					NodeResourceUtilizationThresholds: &api.NodeResourceUtilizationThresholds{
-						Thresholds: test.thresholds,
-					},
-					NodeFit: true,
-				},
-			}
-			HighNodeUtilization(ctx, fakeClient, strategy, nodes, podEvictor)
-
-			podsEvicted := podEvictor.TotalEvicted()
-			if test.expectedPodsEvicted != podsEvicted {
-				t.Errorf("Expected %#v pods to be evicted but %#v got evicted", test.expectedPodsEvicted, podsEvicted)
-			}
-			if evictionFailed {
-				t.Errorf("Pod evictions failed unexpectedly")
-			}
-		})
-	}
-}
-
-func TestValidateHighNodeUtilizationStrategyConfig(t *testing.T) {
-	tests := []struct {
-		name             string
-		thresholds       api.ResourceThresholds
-		targetThresholds api.ResourceThresholds
-		errInfo          error
-	}{
-		{
-			name: "passing target thresholds",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:    20,
-				v1.ResourceMemory: 20,
-			},
-			targetThresholds: api.ResourceThresholds{
-				v1.ResourceCPU:    80,
-				v1.ResourceMemory: 80,
-			},
-			errInfo: fmt.Errorf("targetThresholds is not applicable for HighNodeUtilization"),
-		},
-		{
-			name:       "passing empty thresholds",
-			thresholds: api.ResourceThresholds{},
-			errInfo:    fmt.Errorf("thresholds config is not valid: no resource threshold is configured"),
-		},
-		{
-			name: "passing invalid thresholds",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:    80,
-				v1.ResourceMemory: 120,
-			},
-			errInfo: fmt.Errorf("thresholds config is not valid: %v", fmt.Errorf(
-				"%v threshold not in [%v, %v] range", v1.ResourceMemory, MinResourcePercentage, MaxResourcePercentage)),
-		},
-		{
-			name: "passing valid strategy config",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:    80,
-				v1.ResourceMemory: 80,
-			},
-			errInfo: nil,
-		},
-		{
-			name: "passing valid strategy config with extended resource",
-			thresholds: api.ResourceThresholds{
-				v1.ResourceCPU:    80,
-				v1.ResourceMemory: 80,
-				extendedResource:  80,
-			},
-			errInfo: nil,
-		},
-	}
-
-	for _, testCase := range tests {
-		validateErr := validateHighUtilizationStrategyConfig(testCase.thresholds, testCase.targetThresholds)
-
-		if validateErr == nil || testCase.errInfo == nil {
-			if validateErr != testCase.errInfo {
-				t.Errorf("expected validity of strategy config: thresholds %#v targetThresholds %#v to be %v but got %v instead",
-					testCase.thresholds, testCase.targetThresholds, testCase.errInfo, validateErr)
-			}
-		} else if validateErr.Error() != testCase.errInfo.Error() {
-			t.Errorf("expected validity of strategy config: thresholds %#v targetThresholds %#v to be %v but got %v instead",
-				testCase.thresholds, testCase.targetThresholds, testCase.errInfo, validateErr)
-		}
-	}
-}
-
-func TestHighNodeUtilizationWithTaints(t *testing.T) {
-	ctx := context.Background()
-	strategy := api.DeschedulerStrategy{
-		Enabled: true,
-		Params: &api.StrategyParameters{
-			NodeResourceUtilizationThresholds: &api.NodeResourceUtilizationThresholds{
-				Thresholds: api.ResourceThresholds{
-					v1.ResourceCPU: 40,
-				},
-			},
-		},
-	}
-
-	n1 := test.BuildTestNode("n1", 1000, 3000, 10, nil)
-	n2 := test.BuildTestNode("n2", 1000, 3000, 10, nil)
-	n3 := test.BuildTestNode("n3", 1000, 3000, 10, nil)
-	n3withTaints := n3.DeepCopy()
-	n3withTaints.Spec.Taints = []v1.Taint{
-		{
-			Key:    "key",
-			Value:  "value",
-			Effect: v1.TaintEffectNoSchedule,
-		},
-	}
-
-	podThatToleratesTaint := test.BuildTestPod("tolerate_pod", 200, 0, n1.Name, test.SetRSOwnerRef)
-	podThatToleratesTaint.Spec.Tolerations = []v1.Toleration{
-		{
-			Key:   "key",
-			Value: "value",
-		},
-	}
-
-	tests := []struct {
-		name              string
-		nodes             []*v1.Node
-		pods              []*v1.Pod
-		evictionsExpected int
-	}{
-		{
-			name:  "No taints",
-			nodes: []*v1.Node{n1, n2, n3},
-			pods: []*v1.Pod{
-				//Node 1 pods
-				test.BuildTestPod(fmt.Sprintf("pod_1_%s", n1.Name), 200, 0, n1.Name, test.SetRSOwnerRef),
-				test.BuildTestPod(fmt.Sprintf("pod_2_%s", n1.Name), 200, 0, n1.Name, test.SetRSOwnerRef),
-				test.BuildTestPod(fmt.Sprintf("pod_3_%s", n1.Name), 200, 0, n1.Name, test.SetRSOwnerRef),
-				// Node 2 pods
-				test.BuildTestPod(fmt.Sprintf("pod_4_%s", n2.Name), 200, 0, n2.Name, test.SetRSOwnerRef),
-			},
-			evictionsExpected: 1,
-		},
-		{
-			name:  "No pod tolerates node taint",
-			nodes: []*v1.Node{n1, n3withTaints},
-			pods: []*v1.Pod{
-				//Node 1 pods
-				test.BuildTestPod(fmt.Sprintf("pod_1_%s", n1.Name), 200, 0, n1.Name, test.SetRSOwnerRef),
-				// Node 3 pods
-				test.BuildTestPod(fmt.Sprintf("pod_2_%s", n3withTaints.Name), 200, 0, n3withTaints.Name, test.SetRSOwnerRef),
-			},
-			evictionsExpected: 0,
-		},
-		{
-			name:  "Pod which tolerates node taint",
-			nodes: []*v1.Node{n1, n3withTaints},
-			pods: []*v1.Pod{
-				//Node 1 pods
-				test.BuildTestPod(fmt.Sprintf("pod_1_%s", n1.Name), 100, 0, n1.Name, test.SetRSOwnerRef),
-				podThatToleratesTaint,
-				// Node 3 pods
-				test.BuildTestPod(fmt.Sprintf("pod_9_%s", n3withTaints.Name), 500, 0, n3withTaints.Name, test.SetRSOwnerRef),
-			},
-			evictionsExpected: 1,
-		},
-	}
-
-	for _, item := range tests {
-		t.Run(item.name, func(t *testing.T) {
-			var objs []runtime.Object
-			for _, node := range item.nodes {
-				objs = append(objs, node)
-			}
-
-			for _, pod := range item.pods {
-				objs = append(objs, pod)
-			}
-
-			fakeClient := fake.NewSimpleClientset(objs...)
-
-			podEvictor := evictions.NewPodEvictor(
-				fakeClient,
-				"policy/v1",
-				false,
-				item.evictionsExpected,
-				item.nodes,
-				false,
-				false,
-				false,
-			)
-
-			HighNodeUtilization(ctx, fakeClient, strategy, item.nodes, podEvictor)
-
-			if item.evictionsExpected != podEvictor.TotalEvicted() {
-				t.Errorf("Expected %v evictions, got %v", item.evictionsExpected, podEvictor.TotalEvicted())
-			}
-		})
-	}
-}

pkg/descheduler/strategies/nodeutilization/lownodeutilization.go

@@ -1,188 +0,0 @@
-/*
-Copyright 2017 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package nodeutilization
-
-import (
-	"context"
-	"fmt"
-
-	v1 "k8s.io/api/core/v1"
-	"k8s.io/apimachinery/pkg/api/resource"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	nodeutil "sigs.k8s.io/descheduler/pkg/descheduler/node"
-	"sigs.k8s.io/descheduler/pkg/utils"
-)
-
-// LowNodeUtilization evicts pods from overutilized nodes to underutilized nodes. Note that CPU/Memory requests are used
-// to calculate nodes' utilization and not the actual resource usage.
-func LowNodeUtilization(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, podEvictor *evictions.PodEvictor) {
-	// TODO: May be create a struct for the strategy as well, so that we don't have to pass along the all the params?
-	if err := validateNodeUtilizationParams(strategy.Params); err != nil {
-		klog.ErrorS(err, "Invalid LowNodeUtilization parameters")
-		return
-	}
-	thresholdPriority, err := utils.GetPriorityFromStrategyParams(ctx, client, strategy.Params)
-	if err != nil {
-		klog.ErrorS(err, "Failed to get threshold priority from strategy's params")
-		return
-	}
-
-	nodeFit := false
-	if strategy.Params != nil {
-		nodeFit = strategy.Params.NodeFit
-	}
-
-	thresholds := strategy.Params.NodeResourceUtilizationThresholds.Thresholds
-	targetThresholds := strategy.Params.NodeResourceUtilizationThresholds.TargetThresholds
-	if err := validateLowUtilizationStrategyConfig(thresholds, targetThresholds); err != nil {
-		klog.ErrorS(err, "LowNodeUtilization config is not valid")
-		return
-	}
-	// check if Pods/CPU/Mem are set, if not, set them to 100
-	if _, ok := thresholds[v1.ResourcePods]; !ok {
-		thresholds[v1.ResourcePods] = MaxResourcePercentage
-		targetThresholds[v1.ResourcePods] = MaxResourcePercentage
-	}
-	if _, ok := thresholds[v1.ResourceCPU]; !ok {
-		thresholds[v1.ResourceCPU] = MaxResourcePercentage
-		targetThresholds[v1.ResourceCPU] = MaxResourcePercentage
-	}
-	if _, ok := thresholds[v1.ResourceMemory]; !ok {
-		thresholds[v1.ResourceMemory] = MaxResourcePercentage
-		targetThresholds[v1.ResourceMemory] = MaxResourcePercentage
-	}
-	resourceNames := getResourceNames(thresholds)
-
-	lowNodes, sourceNodes := classifyNodes(
-		getNodeUsage(ctx, client, nodes, thresholds, targetThresholds, resourceNames),
-		// The node has to be schedulable (to be able to move workload there)
-		func(node *v1.Node, usage NodeUsage) bool {
-			if nodeutil.IsNodeUnschedulable(node) {
-				klog.V(2).InfoS("Node is unschedulable, thus not considered as underutilized", "node", klog.KObj(node))
-				return false
-			}
-			return isNodeWithLowUtilization(usage)
-		},
-		func(node *v1.Node, usage NodeUsage) bool {
-			return isNodeAboveTargetUtilization(usage)
-		},
-	)
-
-	// log message in one line
-	keysAndValues := []interface{}{
-		"CPU", thresholds[v1.ResourceCPU],
-		"Mem", thresholds[v1.ResourceMemory],
-		"Pods", thresholds[v1.ResourcePods],
-	}
-	for name := range thresholds {
-		if !isBasicResource(name) {
-			keysAndValues = append(keysAndValues, string(name), int64(thresholds[name]))
-		}
-	}
-	klog.V(1).InfoS("Criteria for a node under utilization", keysAndValues...)
-	klog.V(1).InfoS("Number of underutilized nodes", "totalNumber", len(lowNodes))
-
-	// log message in one line
-	keysAndValues = []interface{}{
-		"CPU", targetThresholds[v1.ResourceCPU],
-		"Mem", targetThresholds[v1.ResourceMemory],
-		"Pods", targetThresholds[v1.ResourcePods],
-	}
-	for name := range targetThresholds {
-		if !isBasicResource(name) {
-			keysAndValues = append(keysAndValues, string(name), int64(targetThresholds[name]))
-		}
-	}
-	klog.V(1).InfoS("Criteria for a node above target utilization", keysAndValues...)
-	klog.V(1).InfoS("Number of overutilized nodes", "totalNumber", len(sourceNodes))
-
-	if len(lowNodes) == 0 {
-		klog.V(1).InfoS("No node is underutilized, nothing to do here, you might tune your thresholds further")
-		return
-	}
-
-	if len(lowNodes) <= strategy.Params.NodeResourceUtilizationThresholds.NumberOfNodes {
-		klog.V(1).InfoS("Number of nodes underutilized is less or equal than NumberOfNodes, nothing to do here", "underutilizedNodes", len(lowNodes), "numberOfNodes", strategy.Params.NodeResourceUtilizationThresholds.NumberOfNodes)
-		return
-	}
-
-	if len(lowNodes) == len(nodes) {
-		klog.V(1).InfoS("All nodes are underutilized, nothing to do here")
-		return
-	}
-
-	if len(sourceNodes) == 0 {
-		klog.V(1).InfoS("All nodes are under target utilization, nothing to do here")
-		return
-	}
-
-	evictable := podEvictor.Evictable(evictions.WithPriorityThreshold(thresholdPriority), evictions.WithNodeFit(nodeFit))
-
-	// stop if node utilization drops below target threshold or any of required capacity (cpu, memory, pods) is moved
-	continueEvictionCond := func(nodeUsage NodeUsage, totalAvailableUsage map[v1.ResourceName]*resource.Quantity) bool {
-		if !isNodeAboveTargetUtilization(nodeUsage) {
-			return false
-		}
-		for name := range totalAvailableUsage {
-			if totalAvailableUsage[name].CmpInt64(0) < 1 {
-				return false
-			}
-		}
-
-		return true
-	}
-
-	evictPodsFromSourceNodes(
-		ctx,
-		sourceNodes,
-		lowNodes,
-		podEvictor,
-		evictable.IsEvictable,
-		resourceNames,
-		"LowNodeUtilization",
-		continueEvictionCond)
-
-	klog.V(1).InfoS("Total number of pods evicted", "evictedPods", podEvictor.TotalEvicted())
-}
-
-// validateLowUtilizationStrategyConfig checks if the strategy's config is valid
-func validateLowUtilizationStrategyConfig(thresholds, targetThresholds api.ResourceThresholds) error {
-	// validate thresholds and targetThresholds config
-	if err := validateThresholds(thresholds); err != nil {
-		return fmt.Errorf("thresholds config is not valid: %v", err)
-	}
-	if err := validateThresholds(targetThresholds); err != nil {
-		return fmt.Errorf("targetThresholds config is not valid: %v", err)
-	}
-
-	// validate if thresholds and targetThresholds have same resources configured
-	if len(thresholds) != len(targetThresholds) {
-		return fmt.Errorf("thresholds and targetThresholds configured different resources")
-	}
-	for resourceName, value := range thresholds {
-		if targetValue, ok := targetThresholds[resourceName]; !ok {
-			return fmt.Errorf("thresholds and targetThresholds configured different resources")
-		} else if value > targetValue {
-			return fmt.Errorf("thresholds' %v percentage is greater than targetThresholds'", resourceName)
-		}
-	}
-	return nil
-}

pkg/descheduler/strategies/nodeutilization/nodeutilization.go

@@ -1,406 +0,0 @@
-/*
-Copyright 2021 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package nodeutilization
-
-import (
-	"context"
-	"fmt"
-	v1 "k8s.io/api/core/v1"
-	"k8s.io/apimachinery/pkg/api/resource"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/utils"
-	"sort"
-)
-
-// NodeUsage stores a node's info, pods on it, thresholds and its resource usage
-type NodeUsage struct {
-	node    *v1.Node
-	usage   map[v1.ResourceName]*resource.Quantity
-	allPods []*v1.Pod
-
-	lowResourceThreshold  map[v1.ResourceName]*resource.Quantity
-	highResourceThreshold map[v1.ResourceName]*resource.Quantity
-}
-
-type continueEvictionCond func(nodeUsage NodeUsage, totalAvailableUsage map[v1.ResourceName]*resource.Quantity) bool
-
-// NodePodsMap is a set of (node, pods) pairs
-type NodePodsMap map[*v1.Node][]*v1.Pod
-
-const (
-	// MinResourcePercentage is the minimum value of a resource's percentage
-	MinResourcePercentage = 0
-	// MaxResourcePercentage is the maximum value of a resource's percentage
-	MaxResourcePercentage = 100
-)
-
-func validateNodeUtilizationParams(params *api.StrategyParameters) error {
-	if params == nil || params.NodeResourceUtilizationThresholds == nil {
-		return fmt.Errorf("NodeResourceUtilizationThresholds not set")
-	}
-	if params.ThresholdPriority != nil && params.ThresholdPriorityClassName != "" {
-		return fmt.Errorf("only one of thresholdPriority and thresholdPriorityClassName can be set")
-	}
-
-	return nil
-}
-
-// validateThresholds checks if thresholds have valid resource name and resource percentage configured
-func validateThresholds(thresholds api.ResourceThresholds) error {
-	if thresholds == nil || len(thresholds) == 0 {
-		return fmt.Errorf("no resource threshold is configured")
-	}
-	for name, percent := range thresholds {
-		if percent < MinResourcePercentage || percent > MaxResourcePercentage {
-			return fmt.Errorf("%v threshold not in [%v, %v] range", name, MinResourcePercentage, MaxResourcePercentage)
-		}
-	}
-	return nil
-}
-
-func getNodeUsage(
-	ctx context.Context,
-	client clientset.Interface,
-	nodes []*v1.Node,
-	lowThreshold, highThreshold api.ResourceThresholds,
-	resourceNames []v1.ResourceName,
-) []NodeUsage {
-	var nodeUsageList []NodeUsage
-
-	for _, node := range nodes {
-		pods, err := podutil.ListPodsOnANode(ctx, client, node)
-		if err != nil {
-			klog.V(2).InfoS("Node will not be processed, error accessing its pods", "node", klog.KObj(node), "err", err)
-			continue
-		}
-
-		// A threshold is in percentages but in <0;100> interval.
-		// Performing `threshold * 0.01` will convert <0;100> interval into <0;1>.
-		// Multiplying it with capacity will give fraction of the capacity corresponding to the given high/low resource threshold in Quantity units.
-		nodeCapacity := node.Status.Capacity
-		if len(node.Status.Allocatable) > 0 {
-			nodeCapacity = node.Status.Allocatable
-		}
-		lowResourceThreshold := map[v1.ResourceName]*resource.Quantity{
-			v1.ResourceCPU:    resource.NewMilliQuantity(int64(float64(lowThreshold[v1.ResourceCPU])*float64(nodeCapacity.Cpu().MilliValue())*0.01), resource.DecimalSI),
-			v1.ResourceMemory: resource.NewQuantity(int64(float64(lowThreshold[v1.ResourceMemory])*float64(nodeCapacity.Memory().Value())*0.01), resource.BinarySI),
-			v1.ResourcePods:   resource.NewQuantity(int64(float64(lowThreshold[v1.ResourcePods])*float64(nodeCapacity.Pods().Value())*0.01), resource.DecimalSI),
-		}
-		for _, name := range resourceNames {
-			if !isBasicResource(name) {
-				cap := nodeCapacity[name]
-				lowResourceThreshold[name] = resource.NewQuantity(int64(float64(lowThreshold[name])*float64(cap.Value())*0.01), resource.DecimalSI)
-			}
-		}
-		highResourceThreshold := map[v1.ResourceName]*resource.Quantity{
-			v1.ResourceCPU:    resource.NewMilliQuantity(int64(float64(highThreshold[v1.ResourceCPU])*float64(nodeCapacity.Cpu().MilliValue())*0.01), resource.DecimalSI),
-			v1.ResourceMemory: resource.NewQuantity(int64(float64(highThreshold[v1.ResourceMemory])*float64(nodeCapacity.Memory().Value())*0.01), resource.BinarySI),
-			v1.ResourcePods:   resource.NewQuantity(int64(float64(highThreshold[v1.ResourcePods])*float64(nodeCapacity.Pods().Value())*0.01), resource.DecimalSI),
-		}
-		for _, name := range resourceNames {
-			if !isBasicResource(name) {
-				cap := nodeCapacity[name]
-				highResourceThreshold[name] = resource.NewQuantity(int64(float64(highThreshold[name])*float64(cap.Value())*0.01), resource.DecimalSI)
-			}
-		}
-
-		nodeUsageList = append(nodeUsageList, NodeUsage{
-			node:                  node,
-			usage:                 nodeUtilization(node, pods, resourceNames),
-			allPods:               pods,
-			lowResourceThreshold:  lowResourceThreshold,
-			highResourceThreshold: highResourceThreshold,
-		})
-	}
-
-	return nodeUsageList
-}
-
-func resourceUsagePercentages(nodeUsage NodeUsage) map[v1.ResourceName]float64 {
-	nodeCapacity := nodeUsage.node.Status.Capacity
-	if len(nodeUsage.node.Status.Allocatable) > 0 {
-		nodeCapacity = nodeUsage.node.Status.Allocatable
-	}
-
-	resourceUsagePercentage := map[v1.ResourceName]float64{}
-	for resourceName, resourceUsage := range nodeUsage.usage {
-		cap := nodeCapacity[resourceName]
-		if !cap.IsZero() {
-			resourceUsagePercentage[resourceName] = 100 * float64(resourceUsage.MilliValue()) / float64(cap.MilliValue())
-		}
-	}
-
-	return resourceUsagePercentage
-}
-
-// classifyNodes classifies the nodes into low-utilization or high-utilization nodes. If a node lies between
-// low and high thresholds, it is simply ignored.
-func classifyNodes(
-	nodeUsages []NodeUsage,
-	lowThresholdFilter, highThresholdFilter func(node *v1.Node, usage NodeUsage) bool,
-) ([]NodeUsage, []NodeUsage) {
-	lowNodes, highNodes := []NodeUsage{}, []NodeUsage{}
-
-	for _, nodeUsage := range nodeUsages {
-		if lowThresholdFilter(nodeUsage.node, nodeUsage) {
-			klog.V(2).InfoS("Node is underutilized", "node", klog.KObj(nodeUsage.node), "usage", nodeUsage.usage, "usagePercentage", resourceUsagePercentages(nodeUsage))
-			lowNodes = append(lowNodes, nodeUsage)
-		} else if highThresholdFilter(nodeUsage.node, nodeUsage) {
-			klog.V(2).InfoS("Node is overutilized", "node", klog.KObj(nodeUsage.node), "usage", nodeUsage.usage, "usagePercentage", resourceUsagePercentages(nodeUsage))
-			highNodes = append(highNodes, nodeUsage)
-		} else {
-			klog.V(2).InfoS("Node is appropriately utilized", "node", klog.KObj(nodeUsage.node), "usage", nodeUsage.usage, "usagePercentage", resourceUsagePercentages(nodeUsage))
-		}
-	}
-
-	return lowNodes, highNodes
-}
-
-// evictPodsFromSourceNodes evicts pods based on priority, if all the pods on the node have priority, if not
-// evicts them based on QoS as fallback option.
-// TODO: @ravig Break this function into smaller functions.
-func evictPodsFromSourceNodes(
-	ctx context.Context,
-	sourceNodes, destinationNodes []NodeUsage,
-	podEvictor *evictions.PodEvictor,
-	podFilter func(pod *v1.Pod) bool,
-	resourceNames []v1.ResourceName,
-	strategy string,
-	continueEviction continueEvictionCond,
-) {
-
-	sortNodesByUsage(sourceNodes)
-
-	// upper bound on total number of pods/cpu/memory and optional extended resources to be moved
-	totalAvailableUsage := map[v1.ResourceName]*resource.Quantity{
-		v1.ResourcePods:   {},
-		v1.ResourceCPU:    {},
-		v1.ResourceMemory: {},
-	}
-
-	var taintsOfDestinationNodes = make(map[string][]v1.Taint, len(destinationNodes))
-	for _, node := range destinationNodes {
-		taintsOfDestinationNodes[node.node.Name] = node.node.Spec.Taints
-
-		for _, name := range resourceNames {
-			if _, ok := totalAvailableUsage[name]; !ok {
-				totalAvailableUsage[name] = resource.NewQuantity(0, resource.DecimalSI)
-			}
-			totalAvailableUsage[name].Add(*node.highResourceThreshold[name])
-			totalAvailableUsage[name].Sub(*node.usage[name])
-		}
-	}
-
-	// log message in one line
-	keysAndValues := []interface{}{
-		"CPU", totalAvailableUsage[v1.ResourceCPU].MilliValue(),
-		"Mem", totalAvailableUsage[v1.ResourceMemory].Value(),
-		"Pods", totalAvailableUsage[v1.ResourcePods].Value(),
-	}
-	for name := range totalAvailableUsage {
-		if !isBasicResource(name) {
-			keysAndValues = append(keysAndValues, string(name), totalAvailableUsage[name].Value())
-		}
-	}
-	klog.V(1).InfoS("Total capacity to be moved", keysAndValues...)
-
-	for _, node := range sourceNodes {
-		klog.V(3).InfoS("Evicting pods from node", "node", klog.KObj(node.node), "usage", node.usage)
-
-		nonRemovablePods, removablePods := classifyPods(node.allPods, podFilter)
-		klog.V(2).InfoS("Pods on node", "node", klog.KObj(node.node), "allPods", len(node.allPods), "nonRemovablePods", len(nonRemovablePods), "removablePods", len(removablePods))
-
-		if len(removablePods) == 0 {
-			klog.V(1).InfoS("No removable pods on node, try next node", "node", klog.KObj(node.node))
-			continue
-		}
-
-		klog.V(1).InfoS("Evicting pods based on priority, if they have same priority, they'll be evicted based on QoS tiers")
-		// sort the evictable Pods based on priority. This also sorts them based on QoS. If there are multiple pods with same priority, they are sorted based on QoS tiers.
-		podutil.SortPodsBasedOnPriorityLowToHigh(removablePods)
-		evictPods(ctx, removablePods, node, totalAvailableUsage, taintsOfDestinationNodes, podEvictor, strategy, continueEviction)
-		klog.V(1).InfoS("Evicted pods from node", "node", klog.KObj(node.node), "evictedPods", podEvictor.NodeEvicted(node.node), "usage", node.usage)
-	}
-}
-
-func evictPods(
-	ctx context.Context,
-	inputPods []*v1.Pod,
-	nodeUsage NodeUsage,
-	totalAvailableUsage map[v1.ResourceName]*resource.Quantity,
-	taintsOfLowNodes map[string][]v1.Taint,
-	podEvictor *evictions.PodEvictor,
-	strategy string,
-	continueEviction continueEvictionCond,
-) {
-
-	if continueEviction(nodeUsage, totalAvailableUsage) {
-		for _, pod := range inputPods {
-			if !utils.PodToleratesTaints(pod, taintsOfLowNodes) {
-				klog.V(3).InfoS("Skipping eviction for pod, doesn't tolerate node taint", "pod", klog.KObj(pod))
-				continue
-			}
-
-			success, err := podEvictor.EvictPod(ctx, pod, nodeUsage.node, strategy)
-			if err != nil {
-				klog.ErrorS(err, "Error evicting pod", "pod", klog.KObj(pod))
-				break
-			}
-
-			if success {
-				klog.V(3).InfoS("Evicted pods", "pod", klog.KObj(pod), "err", err)
-
-				for name := range totalAvailableUsage {
-					if name == v1.ResourcePods {
-						nodeUsage.usage[name].Sub(*resource.NewQuantity(1, resource.DecimalSI))
-						totalAvailableUsage[name].Sub(*resource.NewQuantity(1, resource.DecimalSI))
-					} else {
-						quantity := utils.GetResourceRequestQuantity(pod, name)
-						nodeUsage.usage[name].Sub(quantity)
-						totalAvailableUsage[name].Sub(quantity)
-					}
-				}
-
-				keysAndValues := []interface{}{
-					"node", nodeUsage.node.Name,
-					"CPU", nodeUsage.usage[v1.ResourceCPU].MilliValue(),
-					"Mem", nodeUsage.usage[v1.ResourceMemory].Value(),
-					"Pods", nodeUsage.usage[v1.ResourcePods].Value(),
-				}
-				for name := range totalAvailableUsage {
-					if !isBasicResource(name) {
-						keysAndValues = append(keysAndValues, string(name), totalAvailableUsage[name].Value())
-					}
-				}
-
-				klog.V(3).InfoS("Updated node usage", keysAndValues...)
-				// check if pods can be still evicted
-				if !continueEviction(nodeUsage, totalAvailableUsage) {
-					break
-				}
-			}
-		}
-	}
-}
-
-// sortNodesByUsage sorts nodes based on usage in descending order
-func sortNodesByUsage(nodes []NodeUsage) {
-	sort.Slice(nodes, func(i, j int) bool {
-		ti := nodes[i].usage[v1.ResourceMemory].Value() + nodes[i].usage[v1.ResourceCPU].MilliValue() + nodes[i].usage[v1.ResourcePods].Value()
-		tj := nodes[j].usage[v1.ResourceMemory].Value() + nodes[j].usage[v1.ResourceCPU].MilliValue() + nodes[j].usage[v1.ResourcePods].Value()
-
-		// extended resources
-		for name := range nodes[i].usage {
-			if !isBasicResource(name) {
-				ti = ti + nodes[i].usage[name].Value()
-				tj = tj + nodes[j].usage[name].Value()
-			}
-		}
-
-		// To return sorted in descending order
-		return ti > tj
-	})
-}
-
-// isNodeAboveTargetUtilization checks if a node is overutilized
-// At least one resource has to be above the high threshold
-func isNodeAboveTargetUtilization(usage NodeUsage) bool {
-	for name, nodeValue := range usage.usage {
-		// usage.highResourceThreshold[name] < nodeValue
-		if usage.highResourceThreshold[name].Cmp(*nodeValue) == -1 {
-			return true
-		}
-	}
-	return false
-}
-
-// isNodeWithLowUtilization checks if a node is underutilized
-// All resources have to be below the low threshold
-func isNodeWithLowUtilization(usage NodeUsage) bool {
-	for name, nodeValue := range usage.usage {
-		// usage.lowResourceThreshold[name] < nodeValue
-		if usage.lowResourceThreshold[name].Cmp(*nodeValue) == -1 {
-			return false
-		}
-	}
-
-	return true
-}
-
-// getResourceNames returns list of resource names in resource thresholds
-func getResourceNames(thresholds api.ResourceThresholds) []v1.ResourceName {
-	resourceNames := make([]v1.ResourceName, 0, len(thresholds))
-	for name := range thresholds {
-		resourceNames = append(resourceNames, name)
-	}
-	return resourceNames
-}
-
-// isBasicResource checks if resource is basic native.
-func isBasicResource(name v1.ResourceName) bool {
-	switch name {
-	case v1.ResourceCPU, v1.ResourceMemory, v1.ResourcePods:
-		return true
-	default:
-		return false
-	}
-}
-
-func nodeUtilization(node *v1.Node, pods []*v1.Pod, resourceNames []v1.ResourceName) map[v1.ResourceName]*resource.Quantity {
-	totalReqs := map[v1.ResourceName]*resource.Quantity{
-		v1.ResourceCPU:    resource.NewMilliQuantity(0, resource.DecimalSI),
-		v1.ResourceMemory: resource.NewQuantity(0, resource.BinarySI),
-		v1.ResourcePods:   resource.NewQuantity(int64(len(pods)), resource.DecimalSI),
-	}
-	for _, name := range resourceNames {
-		if !isBasicResource(name) {
-			totalReqs[name] = resource.NewQuantity(0, resource.DecimalSI)
-		}
-	}
-
-	for _, pod := range pods {
-		req, _ := utils.PodRequestsAndLimits(pod)
-		for _, name := range resourceNames {
-			quantity, ok := req[name]
-			if ok && name != v1.ResourcePods {
-				// As Quantity.Add says: Add adds the provided y quantity to the current value. If the current value is zero,
-				// the format of the quantity will be updated to the format of y.
-				totalReqs[name].Add(quantity)
-			}
-		}
-	}
-
-	return totalReqs
-}
-
-func classifyPods(pods []*v1.Pod, filter func(pod *v1.Pod) bool) ([]*v1.Pod, []*v1.Pod) {
-	var nonRemovablePods, removablePods []*v1.Pod
-
-	for _, pod := range pods {
-		if !filter(pod) {
-			nonRemovablePods = append(nonRemovablePods, pod)
-		} else {
-			removablePods = append(removablePods, pod)
-		}
-	}
-
-	return nonRemovablePods, removablePods
-}

pkg/descheduler/strategies/nodeutilization/nodeutilization_test.go

@@ -1,158 +0,0 @@
-/*
-Copyright 2021 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package nodeutilization
-
-import (
-	"fmt"
-	v1 "k8s.io/api/core/v1"
-	"k8s.io/apimachinery/pkg/api/resource"
-	"math"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"testing"
-)
-
-var (
-	lowPriority      = int32(0)
-	highPriority     = int32(10000)
-	extendedResource = v1.ResourceName("example.com/foo")
-)
-
-func TestValidateThresholds(t *testing.T) {
-	tests := []struct {
-		name    string
-		input   api.ResourceThresholds
-		errInfo error
-	}{
-		{
-			name:    "passing nil map for threshold",
-			input:   nil,
-			errInfo: fmt.Errorf("no resource threshold is configured"),
-		},
-		{
-			name:    "passing no threshold",
-			input:   api.ResourceThresholds{},
-			errInfo: fmt.Errorf("no resource threshold is configured"),
-		},
-		{
-			name: "passing extended resource name other than cpu/memory/pods",
-			input: api.ResourceThresholds{
-				v1.ResourceCPU:   40,
-				extendedResource: 50,
-			},
-			errInfo: nil,
-		},
-		{
-			name: "passing invalid resource value",
-			input: api.ResourceThresholds{
-				v1.ResourceCPU:    110,
-				v1.ResourceMemory: 80,
-			},
-			errInfo: fmt.Errorf("%v threshold not in [%v, %v] range", v1.ResourceCPU, MinResourcePercentage, MaxResourcePercentage),
-		},
-		{
-			name: "passing a valid threshold with max and min resource value",
-			input: api.ResourceThresholds{
-				v1.ResourceCPU:    100,
-				v1.ResourceMemory: 0,
-			},
-			errInfo: nil,
-		},
-		{
-			name: "passing a valid threshold with only cpu",
-			input: api.ResourceThresholds{
-				v1.ResourceCPU: 80,
-			},
-			errInfo: nil,
-		},
-		{
-			name: "passing a valid threshold with cpu, memory and pods",
-			input: api.ResourceThresholds{
-				v1.ResourceCPU:    20,
-				v1.ResourceMemory: 30,
-				v1.ResourcePods:   40,
-			},
-			errInfo: nil,
-		},
-		{
-			name: "passing a valid threshold with only extended resource",
-			input: api.ResourceThresholds{
-				extendedResource: 80,
-			},
-			errInfo: nil,
-		},
-		{
-			name: "passing a valid threshold with cpu, memory, pods and extended resource",
-			input: api.ResourceThresholds{
-				v1.ResourceCPU:    20,
-				v1.ResourceMemory: 30,
-				v1.ResourcePods:   40,
-				extendedResource:  50,
-			},
-			errInfo: nil,
-		},
-	}
-
-	for _, test := range tests {
-		validateErr := validateThresholds(test.input)
-
-		if validateErr == nil || test.errInfo == nil {
-			if validateErr != test.errInfo {
-				t.Errorf("expected validity of threshold: %#v to be %v but got %v instead", test.input, test.errInfo, validateErr)
-			}
-		} else if validateErr.Error() != test.errInfo.Error() {
-			t.Errorf("expected validity of threshold: %#v to be %v but got %v instead", test.input, test.errInfo, validateErr)
-		}
-	}
-}
-
-func TestResourceUsagePercentages(t *testing.T) {
-	resourceUsagePercentage := resourceUsagePercentages(NodeUsage{
-		node: &v1.Node{
-			Status: v1.NodeStatus{
-				Capacity: v1.ResourceList{
-					v1.ResourceCPU:    *resource.NewMilliQuantity(2000, resource.DecimalSI),
-					v1.ResourceMemory: *resource.NewQuantity(3977868*1024, resource.BinarySI),
-					v1.ResourcePods:   *resource.NewQuantity(29, resource.BinarySI),
-				},
-				Allocatable: v1.ResourceList{
-					v1.ResourceCPU:    *resource.NewMilliQuantity(1930, resource.DecimalSI),
-					v1.ResourceMemory: *resource.NewQuantity(3287692*1024, resource.BinarySI),
-					v1.ResourcePods:   *resource.NewQuantity(29, resource.BinarySI),
-				},
-			},
-		},
-		usage: map[v1.ResourceName]*resource.Quantity{
-			v1.ResourceCPU:    resource.NewMilliQuantity(1220, resource.DecimalSI),
-			v1.ResourceMemory: resource.NewQuantity(3038982964, resource.BinarySI),
-			v1.ResourcePods:   resource.NewQuantity(11, resource.BinarySI),
-		},
-	})
-
-	expectedUsageInIntPercentage := map[v1.ResourceName]float64{
-		v1.ResourceCPU:    63,
-		v1.ResourceMemory: 90,
-		v1.ResourcePods:   37,
-	}
-
-	for resourceName, percentage := range expectedUsageInIntPercentage {
-		if math.Floor(resourceUsagePercentage[resourceName]) != percentage {
-			t.Errorf("Incorrect percentange computation, expected %v, got math.Floor(%v) instead", percentage, resourceUsagePercentage[resourceName])
-		}
-	}
-
-	t.Logf("resourceUsagePercentage: %#v\n", resourceUsagePercentage)
-}

pkg/descheduler/strategies/pod_antiaffinity_test.go

@@ -1,208 +0,0 @@
-/*
-Copyright 2017 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"testing"
-
-	v1 "k8s.io/api/core/v1"
-	policyv1 "k8s.io/api/policy/v1"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/pkg/utils"
-	"sigs.k8s.io/descheduler/test"
-)
-
-func TestPodAntiAffinity(t *testing.T) {
-	ctx := context.Background()
-	node1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
-	node2 := test.BuildTestNode("n2", 2000, 3000, 10, func(node *v1.Node) {
-		node.ObjectMeta.Labels = map[string]string{
-			"datacenter": "east",
-		}
-	})
-
-	node3 := test.BuildTestNode("n3", 2000, 3000, 10, func(node *v1.Node) {
-		node.Spec = v1.NodeSpec{
-			Unschedulable: true,
-		}
-	})
-
-	p1 := test.BuildTestPod("p1", 100, 0, node1.Name, nil)
-	p2 := test.BuildTestPod("p2", 100, 0, node1.Name, nil)
-	p3 := test.BuildTestPod("p3", 100, 0, node1.Name, nil)
-	p4 := test.BuildTestPod("p4", 100, 0, node1.Name, nil)
-	p5 := test.BuildTestPod("p5", 100, 0, node1.Name, nil)
-	p6 := test.BuildTestPod("p6", 100, 0, node1.Name, nil)
-	p7 := test.BuildTestPod("p7", 100, 0, node1.Name, nil)
-	p8 := test.BuildTestPod("p8", 100, 0, node1.Name, nil)
-
-	criticalPriority := utils.SystemCriticalPriority
-	nonEvictablePod := test.BuildTestPod("non-evict", 100, 0, node1.Name, func(pod *v1.Pod) {
-		pod.Spec.Priority = &criticalPriority
-	})
-	p2.Labels = map[string]string{"foo": "bar"}
-	p5.Labels = map[string]string{"foo": "bar"}
-	p6.Labels = map[string]string{"foo": "bar"}
-	p7.Labels = map[string]string{"foo1": "bar1"}
-	nonEvictablePod.Labels = map[string]string{"foo": "bar"}
-	test.SetNormalOwnerRef(p1)
-	test.SetNormalOwnerRef(p2)
-	test.SetNormalOwnerRef(p3)
-	test.SetNormalOwnerRef(p4)
-	test.SetNormalOwnerRef(p5)
-	test.SetNormalOwnerRef(p6)
-	test.SetNormalOwnerRef(p7)
-
-	// set pod anti affinity
-	setPodAntiAffinity(p1, "foo", "bar")
-	setPodAntiAffinity(p3, "foo", "bar")
-	setPodAntiAffinity(p4, "foo", "bar")
-	setPodAntiAffinity(p5, "foo1", "bar1")
-	setPodAntiAffinity(p6, "foo1", "bar1")
-	setPodAntiAffinity(p7, "foo", "bar")
-
-	// set pod priority
-	test.SetPodPriority(p5, 100)
-	test.SetPodPriority(p6, 50)
-	test.SetPodPriority(p7, 0)
-
-	// Set pod node selectors
-	p8.Spec.NodeSelector = map[string]string{
-		"datacenter": "west",
-	}
-
-	tests := []struct {
-		description             string
-		maxPodsToEvictPerNode   int
-		pods                    []v1.Pod
-		expectedEvictedPodCount int
-		nodeFit                 bool
-		nodes                   []*v1.Node
-	}{
-		{
-			description:             "Maximum pods to evict - 0",
-			maxPodsToEvictPerNode:   0,
-			pods:                    []v1.Pod{*p1, *p2, *p3, *p4},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 3,
-		},
-		{
-			description:             "Maximum pods to evict - 3",
-			maxPodsToEvictPerNode:   3,
-			pods:                    []v1.Pod{*p1, *p2, *p3, *p4},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 3,
-		},
-		{
-			description:             "Evict only 1 pod after sorting",
-			maxPodsToEvictPerNode:   0,
-			pods:                    []v1.Pod{*p5, *p6, *p7},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description:             "Evicts pod that conflicts with critical pod (but does not evict critical pod)",
-			maxPodsToEvictPerNode:   1,
-			pods:                    []v1.Pod{*p1, *nonEvictablePod},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description:             "Evicts pod that conflicts with critical pod (but does not evict critical pod)",
-			maxPodsToEvictPerNode:   1,
-			pods:                    []v1.Pod{*p1, *nonEvictablePod},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description:             "Won't evict pods because node selectors don't match available nodes",
-			maxPodsToEvictPerNode:   1,
-			pods:                    []v1.Pod{*p8, *nonEvictablePod},
-			nodes:                   []*v1.Node{node1, node2},
-			expectedEvictedPodCount: 0,
-			nodeFit:                 true,
-		},
-		{
-			description:             "Won't evict pods because only other node is not schedulable",
-			maxPodsToEvictPerNode:   1,
-			pods:                    []v1.Pod{*p8, *nonEvictablePod},
-			nodes:                   []*v1.Node{node1, node3},
-			expectedEvictedPodCount: 0,
-			nodeFit:                 true,
-		},
-	}
-
-	for _, test := range tests {
-		// create fake client
-		fakeClient := &fake.Clientset{}
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			return true, &v1.PodList{Items: test.pods}, nil
-		})
-		fakeClient.Fake.AddReactor("get", "nodes", func(action core.Action) (bool, runtime.Object, error) {
-			return true, node1, nil
-		})
-
-		podEvictor := evictions.NewPodEvictor(
-			fakeClient,
-			policyv1.SchemeGroupVersion.String(),
-			false,
-			test.maxPodsToEvictPerNode,
-			test.nodes,
-			false,
-			false,
-			false,
-		)
-		strategy := api.DeschedulerStrategy{
-			Params: &api.StrategyParameters{
-				NodeFit: test.nodeFit,
-			},
-		}
-
-		RemovePodsViolatingInterPodAntiAffinity(ctx, fakeClient, strategy, test.nodes, podEvictor)
-		podsEvicted := podEvictor.TotalEvicted()
-		if podsEvicted != test.expectedEvictedPodCount {
-			t.Errorf("Unexpected no of pods evicted: pods evicted: %d, expected: %d", podsEvicted, test.expectedEvictedPodCount)
-		}
-	}
-}
-
-func setPodAntiAffinity(inputPod *v1.Pod, labelKey, labelValue string) {
-	inputPod.Spec.Affinity = &v1.Affinity{
-		PodAntiAffinity: &v1.PodAntiAffinity{
-			RequiredDuringSchedulingIgnoredDuringExecution: []v1.PodAffinityTerm{
-				{
-					LabelSelector: &metav1.LabelSelector{
-						MatchExpressions: []metav1.LabelSelectorRequirement{
-							{
-								Key:      labelKey,
-								Operator: metav1.LabelSelectorOpIn,
-								Values:   []string{labelValue},
-							},
-						},
-					},
-					TopologyKey: "region",
-				},
-			},
-		},
-	}
-}

pkg/descheduler/strategies/pod_lifetime.go

@@ -1,141 +0,0 @@
-/*
-Copyright 2020 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"fmt"
-
-	v1 "k8s.io/api/core/v1"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/utils"
-)
-
-func validatePodLifeTimeParams(params *api.StrategyParameters) error {
-	if params == nil || params.PodLifeTime == nil || params.PodLifeTime.MaxPodLifeTimeSeconds == nil {
-		return fmt.Errorf("MaxPodLifeTimeSeconds not set")
-	}
-
-	if params.PodLifeTime.PodStatusPhases != nil {
-		for _, phase := range params.PodLifeTime.PodStatusPhases {
-			if phase != string(v1.PodPending) && phase != string(v1.PodRunning) {
-				return fmt.Errorf("only Pending and Running phases are supported in PodLifeTime")
-			}
-		}
-	}
-
-	// At most one of include/exclude can be set
-	if params.Namespaces != nil && len(params.Namespaces.Include) > 0 && len(params.Namespaces.Exclude) > 0 {
-		return fmt.Errorf("only one of Include/Exclude namespaces can be set")
-	}
-	if params.ThresholdPriority != nil && params.ThresholdPriorityClassName != "" {
-		return fmt.Errorf("only one of thresholdPriority and thresholdPriorityClassName can be set")
-	}
-
-	return nil
-}
-
-// PodLifeTime evicts pods on nodes that were created more than strategy.Params.MaxPodLifeTimeSeconds seconds ago.
-func PodLifeTime(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, podEvictor *evictions.PodEvictor) {
-	if err := validatePodLifeTimeParams(strategy.Params); err != nil {
-		klog.ErrorS(err, "Invalid PodLifeTime parameters")
-		return
-	}
-
-	thresholdPriority, err := utils.GetPriorityFromStrategyParams(ctx, client, strategy.Params)
-	if err != nil {
-		klog.ErrorS(err, "Failed to get threshold priority from strategy's params")
-		return
-	}
-
-	var includedNamespaces, excludedNamespaces []string
-	if strategy.Params.Namespaces != nil {
-		includedNamespaces = strategy.Params.Namespaces.Include
-		excludedNamespaces = strategy.Params.Namespaces.Exclude
-	}
-
-	evictable := podEvictor.Evictable(evictions.WithPriorityThreshold(thresholdPriority))
-
-	filter := evictable.IsEvictable
-	if strategy.Params.PodLifeTime.PodStatusPhases != nil {
-		filter = func(pod *v1.Pod) bool {
-			for _, phase := range strategy.Params.PodLifeTime.PodStatusPhases {
-				if string(pod.Status.Phase) == phase {
-					return evictable.IsEvictable(pod)
-				}
-			}
-			return false
-		}
-	}
-
-	for _, node := range nodes {
-		klog.V(1).InfoS("Processing node", "node", klog.KObj(node))
-
-		pods := listOldPodsOnNode(ctx, client, node, includedNamespaces, excludedNamespaces, strategy.Params.LabelSelector, *strategy.Params.PodLifeTime.MaxPodLifeTimeSeconds, filter)
-		for _, pod := range pods {
-			success, err := podEvictor.EvictPod(ctx, pod, node, "PodLifeTime")
-			if success {
-				klog.V(1).InfoS("Evicted pod because it exceeded its lifetime", "pod", klog.KObj(pod), "maxPodLifeTime", *strategy.Params.PodLifeTime.MaxPodLifeTimeSeconds)
-			}
-
-			if err != nil {
-				klog.ErrorS(err, "Error evicting pod", "pod", klog.KObj(pod))
-				break
-			}
-		}
-
-	}
-}
-
-func listOldPodsOnNode(
-	ctx context.Context,
-	client clientset.Interface,
-	node *v1.Node,
-	includedNamespaces, excludedNamespaces []string,
-	labelSelector *metav1.LabelSelector,
-	maxPodLifeTimeSeconds uint,
-	filter func(pod *v1.Pod) bool,
-) []*v1.Pod {
-	pods, err := podutil.ListPodsOnANode(
-		ctx,
-		client,
-		node,
-		podutil.WithFilter(filter),
-		podutil.WithNamespaces(includedNamespaces),
-		podutil.WithoutNamespaces(excludedNamespaces),
-		podutil.WithLabelSelector(labelSelector),
-	)
-	if err != nil {
-		return nil
-	}
-
-	var oldPods []*v1.Pod
-	for _, pod := range pods {
-		podAgeSeconds := uint(metav1.Now().Sub(pod.GetCreationTimestamp().Local()).Seconds())
-		if podAgeSeconds > maxPodLifeTimeSeconds {
-			oldPods = append(oldPods, pod)
-		}
-	}
-
-	return oldPods
-}

pkg/descheduler/strategies/pod_lifetime_test.go

@@ -1,276 +0,0 @@
-/*
-Copyright 2020 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"testing"
-	"time"
-
-	v1 "k8s.io/api/core/v1"
-	policyv1 "k8s.io/api/policy/v1"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/test"
-)
-
-func TestPodLifeTime(t *testing.T) {
-	ctx := context.Background()
-	node1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
-	olderPodCreationTime := metav1.NewTime(time.Date(2009, time.November, 10, 23, 0, 0, 0, time.UTC))
-	newerPodCreationTime := metav1.NewTime(time.Now())
-
-	// Setup pods, one should be evicted
-	p1 := test.BuildTestPod("p1", 100, 0, node1.Name, nil)
-	p1.Namespace = "dev"
-	p1.ObjectMeta.CreationTimestamp = newerPodCreationTime
-	p2 := test.BuildTestPod("p2", 100, 0, node1.Name, nil)
-	p2.Namespace = "dev"
-	p2.ObjectMeta.CreationTimestamp = olderPodCreationTime
-
-	ownerRef1 := test.GetReplicaSetOwnerRefList()
-	p1.ObjectMeta.OwnerReferences = ownerRef1
-	p2.ObjectMeta.OwnerReferences = ownerRef1
-
-	// Setup pods, zero should be evicted
-	p3 := test.BuildTestPod("p3", 100, 0, node1.Name, nil)
-	p3.Namespace = "dev"
-	p3.ObjectMeta.CreationTimestamp = newerPodCreationTime
-	p4 := test.BuildTestPod("p4", 100, 0, node1.Name, nil)
-	p4.Namespace = "dev"
-	p4.ObjectMeta.CreationTimestamp = newerPodCreationTime
-
-	ownerRef2 := test.GetReplicaSetOwnerRefList()
-	p3.ObjectMeta.OwnerReferences = ownerRef2
-	p4.ObjectMeta.OwnerReferences = ownerRef2
-
-	// Setup pods, one should be evicted
-	p5 := test.BuildTestPod("p5", 100, 0, node1.Name, nil)
-	p5.Namespace = "dev"
-	p5.ObjectMeta.CreationTimestamp = newerPodCreationTime
-	p6 := test.BuildTestPod("p6", 100, 0, node1.Name, nil)
-	p6.Namespace = "dev"
-	p6.ObjectMeta.CreationTimestamp = metav1.NewTime(time.Now().Add(time.Second * 605))
-
-	ownerRef3 := test.GetReplicaSetOwnerRefList()
-	p5.ObjectMeta.OwnerReferences = ownerRef3
-	p6.ObjectMeta.OwnerReferences = ownerRef3
-
-	// Setup pods, zero should be evicted
-	p7 := test.BuildTestPod("p7", 100, 0, node1.Name, nil)
-	p7.Namespace = "dev"
-	p7.ObjectMeta.CreationTimestamp = newerPodCreationTime
-	p8 := test.BuildTestPod("p8", 100, 0, node1.Name, nil)
-	p8.Namespace = "dev"
-	p8.ObjectMeta.CreationTimestamp = metav1.NewTime(time.Now().Add(time.Second * 595))
-
-	ownerRef4 := test.GetReplicaSetOwnerRefList()
-	p5.ObjectMeta.OwnerReferences = ownerRef4
-	p6.ObjectMeta.OwnerReferences = ownerRef4
-
-	// Setup two old pods with different status phases
-	p9 := test.BuildTestPod("p9", 100, 0, node1.Name, nil)
-	p9.Namespace = "dev"
-	p9.ObjectMeta.CreationTimestamp = olderPodCreationTime
-	p10 := test.BuildTestPod("p10", 100, 0, node1.Name, nil)
-	p10.Namespace = "dev"
-	p10.ObjectMeta.CreationTimestamp = olderPodCreationTime
-
-	p9.Status.Phase = "Pending"
-	p10.Status.Phase = "Running"
-	p9.ObjectMeta.OwnerReferences = ownerRef1
-	p10.ObjectMeta.OwnerReferences = ownerRef1
-
-	p11 := test.BuildTestPod("p11", 100, 0, node1.Name, func(pod *v1.Pod) {
-		pod.Spec.Volumes = []v1.Volume{
-			{
-				Name: "pvc", VolumeSource: v1.VolumeSource{
-					PersistentVolumeClaim: &v1.PersistentVolumeClaimVolumeSource{ClaimName: "foo"},
-				},
-			},
-		}
-		pod.Namespace = "dev"
-		pod.ObjectMeta.CreationTimestamp = olderPodCreationTime
-		pod.ObjectMeta.OwnerReferences = ownerRef1
-	})
-
-	// Setup two old pods with different labels
-	p12 := test.BuildTestPod("p12", 100, 0, node1.Name, nil)
-	p12.Namespace = "dev"
-	p12.ObjectMeta.CreationTimestamp = olderPodCreationTime
-	p13 := test.BuildTestPod("p13", 100, 0, node1.Name, nil)
-	p13.Namespace = "dev"
-	p13.ObjectMeta.CreationTimestamp = olderPodCreationTime
-
-	p12.ObjectMeta.Labels = map[string]string{"foo": "bar"}
-	p13.ObjectMeta.Labels = map[string]string{"foo": "bar1"}
-	p12.ObjectMeta.OwnerReferences = ownerRef1
-	p13.ObjectMeta.OwnerReferences = ownerRef1
-
-	var maxLifeTime uint = 600
-	testCases := []struct {
-		description             string
-		strategy                api.DeschedulerStrategy
-		maxPodsToEvictPerNode   int
-		pods                    []v1.Pod
-		nodes                   []*v1.Node
-		expectedEvictedPodCount int
-		ignorePvcPods           bool
-	}{
-		{
-			description: "Two pods in the `dev` Namespace, 1 is new and 1 very is old. 1 should be evicted.",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p1, *p2},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description: "Two pods in the `dev` Namespace, 2 are new and 0 are old. 0 should be evicted.",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p3, *p4},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 0,
-		},
-		{
-			description: "Two pods in the `dev` Namespace, 1 created 605 seconds ago. 1 should be evicted.",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p5, *p6},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description: "Two pods in the `dev` Namespace, 1 created 595 seconds ago. 0 should be evicted.",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p7, *p8},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 0,
-		},
-		{
-			description: "Two old pods with different status phases. 1 should be evicted.",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{
-						MaxPodLifeTimeSeconds: &maxLifeTime,
-						PodStatusPhases:       []string{"Pending"},
-					},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p9, *p10},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description: "does not evict pvc pods with ignorePvcPods set to true",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p11},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 0,
-			ignorePvcPods:           true,
-		},
-		{
-			description: "evicts pvc pods with ignorePvcPods set to false (or unset)",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p11},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-		{
-			description: "Two old pods with different labels, 1 selected by labelSelector",
-			strategy: api.DeschedulerStrategy{
-				Enabled: true,
-				Params: &api.StrategyParameters{
-					PodLifeTime: &api.PodLifeTime{MaxPodLifeTimeSeconds: &maxLifeTime},
-					LabelSelector: &metav1.LabelSelector{
-						MatchLabels: map[string]string{"foo": "bar"},
-					},
-				},
-			},
-			maxPodsToEvictPerNode:   5,
-			pods:                    []v1.Pod{*p12, *p13},
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-		},
-	}
-
-	for _, tc := range testCases {
-		fakeClient := &fake.Clientset{}
-
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			return true, &v1.PodList{Items: tc.pods}, nil
-		})
-
-		podEvictor := evictions.NewPodEvictor(
-			fakeClient,
-			policyv1.SchemeGroupVersion.String(),
-			false,
-			tc.maxPodsToEvictPerNode,
-			tc.nodes,
-			false,
-			false,
-			tc.ignorePvcPods,
-		)
-
-		PodLifeTime(ctx, fakeClient, tc.strategy, tc.nodes, podEvictor)
-		podsEvicted := podEvictor.TotalEvicted()
-		if podsEvicted != tc.expectedEvictedPodCount {
-			t.Errorf("Test error for description: %s. Expected evicted pods count %v, got %v", tc.description, tc.expectedEvictedPodCount, podsEvicted)
-		}
-	}
-
-}

pkg/descheduler/strategies/toomanyrestarts.go

@@ -1,123 +0,0 @@
-/*
-Copyright 2018 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"fmt"
-
-	v1 "k8s.io/api/core/v1"
-	clientset "k8s.io/client-go/kubernetes"
-	"k8s.io/klog/v2"
-
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
-	"sigs.k8s.io/descheduler/pkg/utils"
-)
-
-func validateRemovePodsHavingTooManyRestartsParams(params *api.StrategyParameters) error {
-	if params == nil || params.PodsHavingTooManyRestarts == nil || params.PodsHavingTooManyRestarts.PodRestartThreshold < 1 {
-		return fmt.Errorf("PodsHavingTooManyRestarts threshold not set")
-	}
-
-	// At most one of include/exclude can be set
-	if params.Namespaces != nil && len(params.Namespaces.Include) > 0 && len(params.Namespaces.Exclude) > 0 {
-		return fmt.Errorf("only one of Include/Exclude namespaces can be set")
-	}
-	if params.ThresholdPriority != nil && params.ThresholdPriorityClassName != "" {
-		return fmt.Errorf("only one of thresholdPriority and thresholdPriorityClassName can be set")
-	}
-
-	return nil
-}
-
-// RemovePodsHavingTooManyRestarts removes the pods that have too many restarts on node.
-// There are too many cases leading this issue: Volume mount failed, app error due to nodes' different settings.
-// As of now, this strategy won't evict daemonsets, mirror pods, critical pods and pods with local storages.
-func RemovePodsHavingTooManyRestarts(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, podEvictor *evictions.PodEvictor) {
-	if err := validateRemovePodsHavingTooManyRestartsParams(strategy.Params); err != nil {
-		klog.ErrorS(err, "Invalid RemovePodsHavingTooManyRestarts parameters")
-		return
-	}
-
-	thresholdPriority, err := utils.GetPriorityFromStrategyParams(ctx, client, strategy.Params)
-	if err != nil {
-		klog.ErrorS(err, "Failed to get threshold priority from strategy's params")
-		return
-	}
-
-	var includedNamespaces, excludedNamespaces []string
-	if strategy.Params.Namespaces != nil {
-		includedNamespaces = strategy.Params.Namespaces.Include
-		excludedNamespaces = strategy.Params.Namespaces.Exclude
-	}
-
-	nodeFit := false
-	if strategy.Params != nil {
-		nodeFit = strategy.Params.NodeFit
-	}
-
-	evictable := podEvictor.Evictable(evictions.WithPriorityThreshold(thresholdPriority), evictions.WithNodeFit(nodeFit))
-
-	for _, node := range nodes {
-		klog.V(1).InfoS("Processing node", "node", klog.KObj(node))
-		pods, err := podutil.ListPodsOnANode(
-			ctx,
-			client,
-			node,
-			podutil.WithFilter(evictable.IsEvictable),
-			podutil.WithNamespaces(includedNamespaces),
-			podutil.WithoutNamespaces(excludedNamespaces),
-			podutil.WithLabelSelector(strategy.Params.LabelSelector),
-		)
-		if err != nil {
-			klog.ErrorS(err, "Error listing a nodes pods", "node", klog.KObj(node))
-			continue
-		}
-
-		for i, pod := range pods {
-			restarts, initRestarts := calcContainerRestarts(pod)
-			if strategy.Params.PodsHavingTooManyRestarts.IncludingInitContainers {
-				if restarts+initRestarts < strategy.Params.PodsHavingTooManyRestarts.PodRestartThreshold {
-					continue
-				}
-			} else if restarts < strategy.Params.PodsHavingTooManyRestarts.PodRestartThreshold {
-				continue
-			}
-			if _, err := podEvictor.EvictPod(ctx, pods[i], node, "TooManyRestarts"); err != nil {
-				klog.ErrorS(err, "Error evicting pod", "pod", klog.KObj(pod))
-				break
-			}
-		}
-	}
-}
-
-// calcContainerRestarts get container restarts and init container restarts.
-func calcContainerRestarts(pod *v1.Pod) (int32, int32) {
-	var restarts, initRestarts int32
-
-	for _, cs := range pod.Status.ContainerStatuses {
-		restarts += cs.RestartCount
-	}
-
-	for _, cs := range pod.Status.InitContainerStatuses {
-		initRestarts += cs.RestartCount
-	}
-
-	return restarts, initRestarts
-}

pkg/descheduler/strategies/toomanyrestarts_test.go

@@ -1,228 +0,0 @@
-/*
-Copyright 2018 The Kubernetes Authors.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License.
-*/
-
-package strategies
-
-import (
-	"context"
-	"testing"
-
-	"fmt"
-
-	v1 "k8s.io/api/core/v1"
-	policyv1 "k8s.io/api/policy/v1"
-	"k8s.io/apimachinery/pkg/api/resource"
-	"k8s.io/apimachinery/pkg/runtime"
-	"k8s.io/client-go/kubernetes/fake"
-	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/api"
-	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
-	"sigs.k8s.io/descheduler/test"
-)
-
-func initPods(node *v1.Node) []v1.Pod {
-	pods := make([]v1.Pod, 0)
-
-	for i := int32(0); i <= 9; i++ {
-		pod := test.BuildTestPod(fmt.Sprintf("pod-%d", i), 100, 0, node.Name, nil)
-		pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-
-		// pod at index i will have 25 * i restarts.
-		pod.Status = v1.PodStatus{
-			InitContainerStatuses: []v1.ContainerStatus{
-				{
-					RestartCount: 5 * i,
-				},
-			},
-			ContainerStatuses: []v1.ContainerStatus{
-				{
-					RestartCount: 10 * i,
-				},
-				{
-					RestartCount: 10 * i,
-				},
-			},
-		}
-		pods = append(pods, *pod)
-	}
-
-	// The following 3 pods won't get evicted.
-	// A daemonset.
-	pods[6].ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
-	// A pod with local storage.
-	pods[7].ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	pods[7].Spec.Volumes = []v1.Volume{
-		{
-			Name: "sample",
-			VolumeSource: v1.VolumeSource{
-				HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-				EmptyDir: &v1.EmptyDirVolumeSource{
-					SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
-			},
-		},
-	}
-	// A Mirror Pod.
-	pods[8].Annotations = test.GetMirrorPodAnnotation()
-
-	return pods
-}
-
-func TestRemovePodsHavingTooManyRestarts(t *testing.T) {
-	ctx := context.Background()
-
-	node1 := test.BuildTestNode("node1", 2000, 3000, 10, nil)
-	node2 := test.BuildTestNode("node2", 2000, 3000, 10, func(node *v1.Node) {
-		node.Spec.Taints = []v1.Taint{
-			{
-				Key:    "hardware",
-				Value:  "gpu",
-				Effect: v1.TaintEffectNoSchedule,
-			},
-		}
-	})
-	node3 := test.BuildTestNode("node3", 2000, 3000, 10, func(node *v1.Node) {
-		node.Spec = v1.NodeSpec{
-			Unschedulable: true,
-		}
-	})
-
-	pods := initPods(node1)
-
-	createStrategy := func(enabled, includingInitContainers bool, restartThresholds int32, nodeFit bool) api.DeschedulerStrategy {
-		return api.DeschedulerStrategy{
-			Enabled: enabled,
-			Params: &api.StrategyParameters{
-				PodsHavingTooManyRestarts: &api.PodsHavingTooManyRestarts{
-					PodRestartThreshold:     restartThresholds,
-					IncludingInitContainers: includingInitContainers,
-				},
-				NodeFit: nodeFit,
-			},
-		}
-	}
-
-	tests := []struct {
-		description             string
-		nodes                   []*v1.Node
-		strategy                api.DeschedulerStrategy
-		expectedEvictedPodCount int
-		maxPodsToEvictPerNode   int
-	}{
-		{
-			description:             "All pods have total restarts under threshold, no pod evictions",
-			strategy:                createStrategy(true, true, 10000, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 0,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Some pods have total restarts bigger than threshold",
-			strategy:                createStrategy(true, true, 1, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 6,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Nine pods have total restarts equals threshold(includingInitContainers=true), 6 pod evictions",
-			strategy:                createStrategy(true, true, 1*25, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 6,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Nine pods have total restarts equals threshold(includingInitContainers=false), 5 pod evictions",
-			strategy:                createStrategy(true, false, 1*25, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 5,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "All pods have total restarts equals threshold(includingInitContainers=true), 6 pod evictions",
-			strategy:                createStrategy(true, true, 1*20, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 6,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Nine pods have total restarts equals threshold(includingInitContainers=false), 6 pod evictions",
-			strategy:                createStrategy(true, false, 1*20, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 6,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Five pods have total restarts bigger than threshold(includingInitContainers=true), but only 1 pod eviction",
-			strategy:                createStrategy(true, true, 5*25+1, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "Five pods have total restarts bigger than threshold(includingInitContainers=false), but only 1 pod eviction",
-			strategy:                createStrategy(true, false, 5*20+1, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 1,
-			maxPodsToEvictPerNode:   0,
-		},
-		{
-			description:             "All pods have total restarts equals threshold(maxPodsToEvictPerNode=3), 3 pod evictions",
-			strategy:                createStrategy(true, true, 1, false),
-			nodes:                   []*v1.Node{node1},
-			expectedEvictedPodCount: 3,
-			maxPodsToEvictPerNode:   3,
-		},
-		{
-			description:             "All pods have total restarts equals threshold(maxPodsToEvictPerNode=3) but the only other node is tained, 0 pod evictions",
-			strategy:                createStrategy(true, true, 1, true),
-			nodes:                   []*v1.Node{node1, node2},
-			expectedEvictedPodCount: 0,
-			maxPodsToEvictPerNode:   3,
-		},
-		{
-			description:             "All pods have total restarts equals threshold(maxPodsToEvictPerNode=3) but the only other node is not schedulable, 0 pod evictions",
-			strategy:                createStrategy(true, true, 1, true),
-			nodes:                   []*v1.Node{node1, node3},
-			expectedEvictedPodCount: 0,
-			maxPodsToEvictPerNode:   3,
-		},
-	}
-
-	for _, tc := range tests {
-
-		fakeClient := &fake.Clientset{}
-		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-			return true, &v1.PodList{Items: pods}, nil
-		})
-
-		podEvictor := evictions.NewPodEvictor(
-			fakeClient,
-			policyv1.SchemeGroupVersion.String(),
-			false,
-			tc.maxPodsToEvictPerNode,
-			tc.nodes,
-			false,
-			false,
-			false,
-		)
-
-		RemovePodsHavingTooManyRestarts(ctx, fakeClient, tc.strategy, tc.nodes, podEvictor)
-		actualEvictedPodCount := podEvictor.TotalEvicted()
-		if actualEvictedPodCount != tc.expectedEvictedPodCount {
-			t.Errorf("Test %#v failed, expected %v pod evictions, but got %v pod evictions\n", tc.description, tc.expectedEvictedPodCount, actualEvictedPodCount)
-		}
-	}
-
-}