Merge pull request #354 from damemi/bump-chart

Bump Helm chart to version 0.18.1

.github/workflows/release.yaml

@@ -0,0 +1,30 @@
+name: Release Charts
+
+on:
+  push:
+    tags:
+      - chart-*
+
+jobs:
+  release:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Configure Git
+        run: |
+          git config user.name "$GITHUB_ACTOR"
+          git config user.email "$GITHUB_ACTOR@users.noreply.github.com"
+
+      - name: Fetch history
+        run: git fetch --prune --unshallow
+
+      - name: Add dependency chart repos
+        run: |
+          helm repo add stable https://kubernetes-charts.storage.googleapis.com/
+
+      - name: Run chart-releaser
+        uses: helm/chart-releaser-action@v1.0.0-rc.2
+        env:
+          CR_TOKEN: "${{ secrets.GITHUB_TOKEN }}"

.gitignore

@@ -1 +1,5 @@
 _output/
+_tmp/
+vendordiff.patch
+.idea/
+*.code-workspace

.travis.yml

@@ -1,31 +0,0 @@
-sudo: false
-
-language: go
-
-go:
-  - 1.13.x
-env:
-  - K8S_VERSION=v1.17.0
-  - K8S_VERSION=v1.16.4
-  - K8S_VERSION=v1.15.7
-services:
-  - docker
-before_script:
-  - curl -Lo kubectl https://storage.googleapis.com/kubernetes-release/release/${K8S_VERSION}/bin/linux/amd64/kubectl && chmod +x kubectl && sudo mv kubectl /usr/local/bin/
-  - wget https://github.com/kubernetes-sigs/kind/releases/download/v0.7.0/kind-linux-amd64
-  - chmod +x kind-linux-amd64
-  - mv kind-linux-amd64 kind
-  - export PATH=$PATH:$PWD
-  - kind create cluster --image kindest/node:${K8S_VERSION} --config=$TRAVIS_BUILD_DIR/hack/kind_config.yaml
-  - export KUBECONFIG="$(kind get kubeconfig-path)"
-  - docker pull kubernetes/pause
-  - kind load docker-image kubernetes/pause
-  - kind get kubeconfig > /tmp/admin.conf
-script:
-- mkdir -p ~/gopath/src/sigs.k8s.io/
-- mv ~/gopath/src/github.com/kubernetes-sigs/descheduler ~/gopath/src/sigs.k8s.io/.
-- hack/verify-gofmt.sh
-- make lint
-- make build
-- make test-unit
-- make test-e2e

CONTRIBUTING.md

@@ -10,7 +10,7 @@ We have full documentation on how to get started contributing here:
 
 - [Contributor License Agreement](https://git.k8s.io/community/CLA.md) Kubernetes projects require that you sign a Contributor License Agreement (CLA) before we can accept your pull requests
 - [Kubernetes Contributor Guide](http://git.k8s.io/community/contributors/guide) - Main contributor documentation, or you can just jump directly to the [contributing section](http://git.k8s.io/community/contributors/guide#contributing)
-- [Contributor Cheat Sheet](https://git.k8s.io/community/contributors/guide/contributor-cheatsheet.md) - Common resources for existing developers
+- [Contributor Cheat Sheet](https://git.k8s.io/community/contributors/guide/contributor-cheatsheet/README.md) - Common resources for existing developers
 
 ## Mentorship
 

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:1.13.6
+FROM golang:1.13.9
 
 WORKDIR /go/src/sigs.k8s.io/descheduler
 COPY . .
@@ -19,7 +19,7 @@ RUN make
 
 FROM scratch
 
-MAINTAINER Avesh Agarwal <avagarwa@redhat.com>
+MAINTAINER Avesh Agarwal <avesh.ncsu@gmail.com>
 
 COPY --from=0 /go/src/sigs.k8s.io/descheduler/_output/bin/descheduler /bin/descheduler
 

Dockerfile.dev

@@ -13,7 +13,7 @@
 # limitations under the License.
 FROM scratch
 
-MAINTAINER Avesh Agarwal <avagarwa@redhat.com>
+MAINTAINER Avesh Agarwal <avesh.ncsu@gmail.com>
 
 COPY _output/bin/descheduler /bin/descheduler
 

Makefile

@@ -41,6 +41,8 @@ IMAGE_GCLOUD:=$(REGISTRY)/descheduler:$(VERSION)
 # In the future binaries can be uploaded to
 # GCS bucket gs://k8s-staging-descheduler.
 
+HAS_HELM := $(shell which helm)
+
 all: build
 
 build:
@@ -62,6 +64,14 @@ push: push-container-to-gcloud
 clean:
 	rm -rf _output
 
+verify: verify-gofmt verify-vendor lint lint-chart
+
+verify-gofmt:
+	./hack/verify-gofmt.sh
+
+verify-vendor:
+	./hack/verify-vendor.sh
+
 test-unit:
 	./test/run-unit-tests.sh
 
@@ -76,6 +86,12 @@ gen:
 	go mod tidy
 lint:
 ifndef HAS_GOLANGCI
-	curl -sfL https://install.goreleaser.com/github.com/golangci/golangci-lint.sh | sh -s -- -b $(GOPATH)/bin ${GOLANGCI_VERSION}
+	curl -sfL https://install.goreleaser.com/github.com/golangci/golangci-lint.sh | sh -s -- -b ./_output/bin ${GOLANGCI_VERSION}
 endif
-	golangci-lint run
+	./_output/bin/golangci-lint run
+
+lint-chart:
+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
+	helm lint ./charts/descheduler

OWNERS

@@ -2,6 +2,7 @@ approvers:
 - aveshagarwal
 - k82cn
 - ravisantoshgudimetla
+- damemi
 reviewers:
 - aveshagarwal
 - k82cn

README.md

@@ -1,4 +1,3 @@
-[![Build Status](https://travis-ci.org/kubernetes-sigs/descheduler.svg?branch=master)](https://travis-ci.org/kubernetes-sigs/descheduler)
 [![Go Report Card](https://goreportcard.com/badge/kubernetes-sigs/descheduler)](https://goreportcard.com/report/sigs.k8s.io/descheduler)
 
 # Descheduler for Kubernetes
@@ -9,8 +8,8 @@ Scheduling in Kubernetes is the process of binding pending pods to nodes, and is
 a component of Kubernetes called kube-scheduler. The scheduler's decisions, whether or where a
 pod can or can not be scheduled, are guided by its configurable policy which comprises of set of
 rules, called predicates and priorities. The scheduler's decisions are influenced by its view of
-a Kubernetes cluster at that point of time when a new pod appears first time for scheduling.
-As Kubernetes clusters are very dynamic and their state change over time, there may be desired
+a Kubernetes cluster at that point of time when a new pod appears for scheduling.
+As Kubernetes clusters are very dynamic and their state changes over time, there may be desire
 to move already running pods to some other nodes for various reasons:
 
 * Some nodes are under or over utilized.
@@ -24,87 +23,63 @@ Descheduler, based on its policy, finds pods that can be moved and evicts them.
 note, in current implementation, descheduler does not schedule replacement of evicted pods
 but relies on the default scheduler for that.
 
-## Build and Run
+## Quick Start
 
-- Checkout the repo into your $GOPATH directory under src/sigs.k8s.io/descheduler
+The descheduler can be run as a Job or CronJob inside of a k8s cluster. It has the
+advantage of being able to be run multiple times without needing user intervention.
+The descheduler pod is run as a critical pod in the `kube-system` namespace to avoid
+being evicted by itself or by the kubelet.
 
-Build descheduler:
-
-```sh
-$ make
-```
-
-and run descheduler:
-
-```sh
-$ ./_output/bin/descheduler --kubeconfig <path to kubeconfig> --policy-config-file <path-to-policy-file>
-```
-
-If you want more information about what descheduler is doing add `-v 1` to the command line
-
-For more information about available options run:
-```
-$ ./_output/bin/descheduler --help
-```
-
-## Running Descheduler as a Job or CronJob
-
-The descheduler can be run as a job or cronjob inside of a pod. It has the advantage of
-being able to be run multiple times without needing user intervention.
-The descheduler pod is run as a critical pod to avoid being evicted by itself,
-or by the kubelet due to an eviction event. Since critical pods are created in the
-`kube-system` namespace, the descheduler job and its pod will also be created
-in `kube-system` namespace.
-
-### Setup RBAC
-
-To give necessary permissions for the descheduler to work in a pod.
+### Run As A Job
 
 ```
-$ kubectl create -f kubernetes/rbac.yaml
+kubectl create -f kubernetes/rbac.yaml
+kubectl create -f kubernetes/configmap.yaml
+kubectl create -f kubernetes/job.yaml
 ```
 
-### Create a configmap to store descheduler policy
+### Run As A CronJob
 
 ```
-$ kubectl create -f kubernetes/configmap.yaml
+kubectl create -f kubernetes/rbac.yaml
+kubectl create -f kubernetes/configmap.yaml
+kubectl create -f kubernetes/cronjob.yaml
 ```
 
-### Create a Job or CronJob
+## User Guide
 
-As a Job.
-```
-$ kubectl create -f kubernetes/job.yaml
-```
-
-Or as a CronJob.
-```
-$ kubectl create -f kubernetes/cronjob.yaml
-```
+See the [user guide](docs/user-guide.md) in the `/docs` directory.
 
 ## Policy and Strategies
 
-Descheduler's policy is configurable and includes strategies to be enabled or disabled.
-Five strategies, `RemoveDuplicates`, `LowNodeUtilization`, `RemovePodsViolatingInterPodAntiAffinity`, `RemovePodsViolatingNodeAffinity` , `RemovePodsViolatingNodeTaints` are currently implemented.
-As part of the policy, the parameters associated with the strategies can be configured too.
+Descheduler's policy is configurable and includes strategies that can be enabled or disabled.
+Seven strategies `RemoveDuplicates`, `LowNodeUtilization`, `RemovePodsViolatingInterPodAntiAffinity`,
+`RemovePodsViolatingNodeAffinity`, `RemovePodsViolatingNodeTaints`, `RemovePodsHavingTooManyRestarts`, and `PodLifeTime`
+are currently implemented. As part of the policy, the parameters associated with the strategies can be configured too.
 By default, all strategies are enabled.
 
 ### RemoveDuplicates
 
 This strategy makes sure that there is only one pod associated with a Replica Set (RS),
-Replication Controller (RC), Deployment, or Job running on same node. If there are more,
+Replication Controller (RC), Deployment, or Job running on the same node. If there are more,
 those duplicate pods are evicted for better spreading of pods in a cluster. This issue could happen
 if some nodes went down due to whatever reasons, and pods on them were moved to other nodes leading to
-more than one pod associated with RS or RC, for example, running on same node. Once the failed nodes
-are ready again, this strategy could be enabled to evict those duplicate pods. Currently, there are no
-parameters associated with this strategy. To disable this strategy, the policy should look like:
+more than one pod associated with a RS or RC, for example, running on the same node. Once the failed nodes
+are ready again, this strategy could be enabled to evict those duplicate pods.
+
+It provides one optional parameter, `ExcludeOwnerKinds`, which is a list of OwnerRef `Kind`s. If a pod
+has any of these `Kind`s listed as an `OwnerRef`, that pod will not be considered for eviction.
 
 ```
 apiVersion: "descheduler/v1alpha1"
 kind: "DeschedulerPolicy"
 strategies:
   "RemoveDuplicates":
-     enabled: false
+     enabled: true
+     params:
+       removeDuplicates:
+         excludeOwnerKinds:
+         - "ReplicaSet"
 ```
 
 ### LowNodeUtilization
@@ -113,10 +88,10 @@ This strategy finds nodes that are under utilized and evicts pods, if possible,
 in the hope that recreation of evicted pods will be scheduled on these underutilized nodes. The
 parameters of this strategy are configured under `nodeResourceUtilizationThresholds`.
 
-The under utilization of nodes is determined by a configurable threshold, `thresholds`. The threshold
+The under utilization of nodes is determined by a configurable threshold `thresholds`. The threshold
 `thresholds` can be configured for cpu, memory, and number of pods in terms of percentage. If a node's
 usage is below threshold for all (cpu, memory, and number of pods), the node is considered underutilized.
-Currently, pods' request resource requirements are considered for computing node resource utilization.
+Currently, pods request resource requirements are considered for computing node resource utilization.
 
 There is another configurable threshold, `targetThresholds`, that is used to compute those potential nodes
 from where pods could be evicted. Any node, between the thresholds, `thresholds` and `targetThresholds` is
@@ -124,7 +99,7 @@ considered appropriately utilized and is not considered for eviction. The thresh
 can be configured for cpu, memory, and number of pods too in terms of percentage.
 
 These thresholds, `thresholds` and `targetThresholds`, could be tuned as per your cluster requirements.
-An example of the policy for this strategy would look like:
+Here is an example of a policy for this strategy:
 
 ```
 apiVersion: "descheduler/v1alpha1"
@@ -144,14 +119,19 @@ strategies:
            "pods": 50
 ```
 
-There is another parameter associated with `LowNodeUtilization` strategy, called `numberOfNodes`.
-This parameter can be configured to activate the strategy only when number of under utilized nodes
+There is another parameter associated with the `LowNodeUtilization` strategy, called `numberOfNodes`.
+This parameter can be configured to activate the strategy only when the number of under utilized nodes
 are above the configured value. This could be helpful in large clusters where a few nodes could go
 under utilized frequently or for a short period of time. By default, `numberOfNodes` is set to zero.
 
 ### RemovePodsViolatingInterPodAntiAffinity
 
-This strategy makes sure that pods violating interpod anti-affinity are removed from nodes. For example, if there is podA on node and podB and podC(running on same node) have antiaffinity rules which prohibit them to run on the same node, then podA will be evicted from the node so that podB and podC could run. This issue could happen, when the anti-affinity rules for pods B,C are created when they are already running on node. Currently, there are no parameters associated with this strategy. To disable this strategy, the policy should look like:
+This strategy makes sure that pods violating interpod anti-affinity are removed from nodes. For example,
+if there is podA on a node and podB and podC (running on the same node) have anti-affinity rules which prohibit
+them to run on the same node, then podA will be evicted from the node so that podB and podC could run. This
+issue could happen, when the anti-affinity rules for podB and podC are created when they are already running on
+node. Currently, there are no parameters associated with this strategy. To disable this strategy, the
+policy should look like:
 
 ```
 apiVersion: "descheduler/v1alpha1"
@@ -163,7 +143,23 @@ strategies:
 
 ### RemovePodsViolatingNodeAffinity
 
-This strategy makes sure that pods violating node affinity are removed from nodes. For example, there is podA that was scheduled on nodeA which satisfied the node affinity rule `requiredDuringSchedulingIgnoredDuringExecution` at the time of scheduling, but over time nodeA no longer satisfies the rule, then if another node nodeB is available that satisfies the node affinity rule, then podA will be evicted from nodeA. The policy file should like this -
+This strategy makes sure all pods violating
+[node affinity](https://kubernetes.io/docs/concepts/configuration/assign-pod-node/#node-affinity)
+are eventually removed from nodes. Node affinity rules allow a pod to specify
+`requiredDuringSchedulingIgnoredDuringExecution` type, which tells the scheduler
+to respect node affinity when scheduling the pod but kubelet to ignore
+in case node changes over time and no longer respects the affinity.
+When enabled, the strategy serves as a temporary implementation
+of `requiredDuringSchedulingRequiredDuringExecution` and evicts pod for kubelet
+that no longer respects node affinity.
+
+For example, there is podA scheduled on nodeA which satisfies the node
+affinity rule `requiredDuringSchedulingIgnoredDuringExecution` at the time
+of scheduling. Over time nodeA stops to satisfy the rule. When the strategy gets
+executed and there is another node available that satisfies the node affinity rule,
+podA gets evicted from nodeA.
+
+The policy file should look like:
 
 ```
 apiVersion: "descheduler/v1alpha1"
@@ -175,9 +171,13 @@ strategies:
       nodeAffinityType:
       - "requiredDuringSchedulingIgnoredDuringExecution"
 ```
+
 ### RemovePodsViolatingNodeTaints
 
-This strategy makes sure that pods violating NoSchedule taints on nodes are removed. For example: there is a pod "podA" with toleration to tolerate a taint ``key=value:NoSchedule`` scheduled and running on the tainted node. If the node's taint is subsequently updated/removed, taint is no longer satisfied by its pods' tolerations and will be evicted. The policy file should look like:
+This strategy makes sure that pods violating NoSchedule taints on nodes are removed. For example there is a
+pod "podA" with a toleration to tolerate a taint ``key=value:NoSchedule`` scheduled and running on the tainted
+node. If the node's taint is subsequently updated/removed, taint is no longer satisfied by its pods' tolerations
+and will be evicted. The policy file should look like:
 
 ````
 apiVersion: "descheduler/v1alpha1"
@@ -186,52 +186,119 @@ strategies:
   "RemovePodsViolatingNodeTaints":
     enabled: true
 ````
+
+### RemovePodsHavingTooManyRestarts
+
+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.
+
+```
+apiVersion: "descheduler/v1alpha1"
+kind: "DeschedulerPolicy"
+strategies:
+  "RemovePodsHavingTooManyRestarts":
+     enabled: true
+     params:
+       podsHavingTooManyRestarts:
+         podRestartThreshold: 100
+         includingInitContainers: true
+```
+
+### PodLifeTime
+
+This strategy evicts pods that are older than `.strategies.PodLifeTime.params.maxPodLifeTimeSeconds` The policy
+file should look like:
+
+````
+apiVersion: "descheduler/v1alpha1"
+kind: "DeschedulerPolicy"
+strategies:
+  "PodLifeTime":
+     enabled: true
+     params:
+        maxPodLifeTimeSeconds: 86400
+````
+
 ## Pod Evictions
 
-When the descheduler decides to evict pods from a node, it employs following general mechanism:
+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.
-* Pods (static or mirrored pods or stand alone pods) not part of an RC, RS, Deployment or Jobs are
+* Pods (static or mirrored pods or stand alone pods) not part of an RC, RS, Deployment or Job are
 never evicted because these pods won't be recreated.
 * Pods associated with DaemonSets are never evicted.
 * Pods with local storage are never evicted.
-* Best efforts pods are evicted before Burstable and Guaranteed pods.
-* All types of pods with annotation descheduler.alpha.kubernetes.io/evict are evicted. This 
-annotation is used to override checks which prevent eviction and user can select which pod is evicted. 
-User should know how and if the pod will be recreated.
+* Best efforts pods are evicted before burstable and guaranteed pods.
+* All types of pods with the annotation descheduler.alpha.kubernetes.io/evict are evicted. 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.
 
-### Pod disruption Budget (PDB)
-Pods subject to Pod Disruption Budget (PDB) are not evicted if descheduling violates its pod
-disruption budget (PDB). The pods are evicted by using eviction subresource to handle PDB.
+### Pod Disruption Budget (PDB)
+
+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.
+
+## Compatibility Matrix
+The below compatibility matrix shows the k8s client package(client-go, apimachinery, etc) versions that descheduler
+is compiled with. At this time descheduler does not have a hard dependency to a specific k8s release. However a
+particular descheduler release is only tested against the three latest k8s minor versions. For example descheduler
+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
+-------------|-----------------------------
+v0.18        | v1.18
+v0.10        | v1.17
+v0.4-v0.9    | v1.9+
+v0.1-v0.3    | v1.7-v1.8
+
+
+## Getting Involved and Contributing
+
+Are you interested in contributing to descheduler? We, the
+maintainers and community, would love your suggestions, contributions, and help!
+Also, the maintainers can be contacted at any time to learn more about how to get
+involved.
+
+To get started writing code see the [contributor guide](docs/contributor-guide.md) in the `/docs` directory.
+
+In the interest of getting more new people involved we tag issues with
+[`good first issue`][good_first_issue].
+These are typically issues that have smaller scope but are good ways to start
+to get acquainted with the codebase.
+
+We also encourage ALL active community participants to act as if they are
+maintainers, even if you don't have "official" write permissions. This is a
+community effort, we are here to serve the Kubernetes community. If you have an
+active interest and you want to get involved, you have real power! Don't assume
+that the only people who can get things done around here are the "maintainers".
+
+We also would love to add more "official" maintainers, so show us what you can
+do!
+
+This repository uses the Kubernetes bots. See a full list of the commands [here][prow].
+
+### Communicating With Contributors
+
+You can reach the contributors of this project at:
+
+- [Slack channel](https://kubernetes.slack.com/messages/sig-scheduling)
+- [Mailing list](https://groups.google.com/forum/#!forum/kubernetes-sig-scheduling)
+
+Learn how to engage with the Kubernetes community on the [community page](http://kubernetes.io/community/).
 
 ## Roadmap
 
 This roadmap is not in any particular order.
 
 * Consideration of pod affinity
-* Strategy to consider pod life time
 * Strategy to consider number of pending pods
 * Integration with cluster autoscaler
 * Integration with metrics providers for obtaining real load metrics
 * Consideration of Kubernetes's scheduler's predicates
 
 
-## Compatibility matrix
-
-Descheduler  | supported Kubernetes version
--------------|-----------------------------
-0.4+          | 1.9+
-0.1-0.3      | 1.7-1.8
-
-## Community, discussion, contribution, and support
-
-Learn how to engage with the Kubernetes community on the [community page](http://kubernetes.io/community/).
-
-You can reach the maintainers of this project at:
-
-- [Slack channel](https://kubernetes.slack.com/messages/sig-scheduling)
-- [Mailing list](https://groups.google.com/forum/#!forum/kubernetes-sig-scheduling)
-
 ### Code of conduct
 
 Participation in the Kubernetes community is governed by the [Kubernetes Code of Conduct](code-of-conduct.md).

charts/descheduler/.helmignore

@@ -0,0 +1,22 @@
+# Patterns to ignore when building packages.
+# This supports shell glob matching, relative path matching, and
+# negation (prefixed with !). Only one pattern per line.
+.DS_Store
+# Common VCS dirs
+.git/
+.gitignore
+.bzr/
+.bzrignore
+.hg/
+.hgignore
+.svn/
+# Common backup files
+*.swp
+*.bak
+*.tmp
+*~
+# Various IDEs
+.project
+.idea/
+*.tmproj
+.vscode/

charts/descheduler/Chart.yaml

@@ -0,0 +1,16 @@
+apiVersion: v1
+name: descheduler
+version: 0.18.1
+appVersion: 0.18.0
+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
+- descheduler
+- kube-scheduler
+home: https://github.com/kubernetes-sigs/descheduler
+icon: https://kubernetes.io/images/favicon.png
+sources:
+- https://github.com/kubernetes-sigs/descheduler
+maintainers:
+- name: stevehipwell
+  email: steve.hipwell@github.com

charts/descheduler/README.md

@@ -0,0 +1,59 @@
+# Descheduler for Kubernetes
+
+[Descheduler](https://github.com/kubernetes-sigs/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.
+
+## TL;DR:
+
+```shell
+helm repo add descheduler https://kubernetes-sigs.github.io/descheduler/
+$ helm install descheduler/descheduler --name my-release
+```
+
+## Introduction
+
+This chart bootstraps a [desheduler](https://github.com/kubernetes-sigs/descheduler/) cron job on a [Kubernetes](http://kubernetes.io) cluster using the [Helm](https://helm.sh) package manager.
+
+## Prerequisites
+
+- Kubernetes 1.14+
+
+## Installing the Chart
+
+To install the chart with the release name `my-release`:
+
+```shell
+helm install --name my-release descheduler/descheduler
+```
+
+The command deploys _descheduler_ on the Kubernetes cluster in the default configuration. The [configuration](#configuration) section lists the parameters that can be configured during installation.
+
+> **Tip**: List all releases using `helm list`
+
+## Uninstalling the Chart
+
+To uninstall/delete the `my-release` deployment:
+
+```shell
+helm delete my-release
+```
+
+The command removes all the Kubernetes components associated with the chart and deletes the release.
+
+## Configuration
+
+The following table lists the configurable parameters of the _descheduler_ chart and their default values.
+
+| Parameter                      | Description                                                                                                           | Default                                                |
+| ------------------------------ | --------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------ |
+| `image.repository`             | Docker repository to use                                                                                              | `us.gcr.io/k8s-artifacts-prod/descheduler/descheduler` |
+| `image.tag`                    | Docker tag to use                                                                                                     | `v[chart appVersion]`                                              |
+| `image.pullPolicy`             | Docker image pull policy                                                                                              | `IfNotPresent`                                         |
+| `nameOverride`                 | String to partially override `descheduler.fullname` template (will prepend the release name)                          | `""`                                                   |
+| `fullnameOverride`             | String to fully override `descheduler.fullname` template                                                              | `""`                                                   |
+| `schedule`                     | The cron schedule to run the _descheduler_ job on                                                                     | `"*/2 * * * *"`                                        |
+| `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`                                                 |
+| `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`                                                  |

charts/descheduler/templates/NOTES.txt

@@ -0,0 +1 @@
+Descheduler installed as a cron job.

charts/descheduler/templates/_helpers.tpl

@@ -0,0 +1,56 @@
+{{/* vim: set filetype=mustache: */}}
+{{/*
+Expand the name of the chart.
+*/}}
+{{- define "descheduler.name" -}}
+{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" -}}
+{{- end -}}
+
+{{/*
+Create a default fully qualified app name.
+We truncate at 63 chars because some Kubernetes name fields are limited to this (by the DNS naming spec).
+If release name contains chart name it will be used as a full name.
+*/}}
+{{- define "descheduler.fullname" -}}
+{{- if .Values.fullnameOverride -}}
+{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" -}}
+{{- else -}}
+{{- $name := default .Chart.Name .Values.nameOverride -}}
+{{- if contains $name .Release.Name -}}
+{{- .Release.Name | trunc 63 | trimSuffix "-" -}}
+{{- else -}}
+{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" -}}
+{{- end -}}
+{{- end -}}
+{{- end -}}
+
+{{/*
+Create chart name and version as used by the chart label.
+*/}}
+{{- define "descheduler.chart" -}}
+{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" -}}
+{{- end -}}
+
+{{/*
+Common labels
+*/}}
+{{- define "descheduler.labels" -}}
+app.kubernetes.io/name: {{ include "descheduler.name" . }}
+helm.sh/chart: {{ include "descheduler.chart" . }}
+app.kubernetes.io/instance: {{ .Release.Name }}
+{{- if .Chart.AppVersion }}
+app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
+{{- end }}
+app.kubernetes.io/managed-by: {{ .Release.Service }}
+{{- end -}}
+
+{{/*
+Create the name of the service account to use
+*/}}
+{{- define "descheduler.serviceAccountName" -}}
+{{- if .Values.serviceAccount.create -}}
+    {{ default (include "descheduler.fullname" .) .Values.serviceAccount.name }}
+{{- else -}}
+    {{ default "default" .Values.serviceAccount.name }}
+{{- end -}}
+{{- end -}}

charts/descheduler/templates/clusterrole.yaml

@@ -0,0 +1,21 @@
+{{- if .Values.rbac.create -}}
+kind: ClusterRole
+apiVersion: rbac.authorization.k8s.io/v1
+metadata:
+  name: {{ template "descheduler.fullname" . }}
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+rules:
+- apiGroups: [""]
+  resources: ["events"]
+  verbs: ["create", "update"]
+- apiGroups: [""]
+  resources: ["nodes"]
+  verbs: ["get", "watch", "list"]
+- apiGroups: [""]
+  resources: ["pods"]
+  verbs: ["get", "watch", "list", "delete"]
+- apiGroups: [""]
+  resources: ["pods/eviction"]
+  verbs: ["create"]
+{{- end -}}

charts/descheduler/templates/clusterrolebinding.yaml

@@ -0,0 +1,16 @@
+{{- if .Values.rbac.create -}}
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: {{ template "descheduler.fullname" . }}
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: {{ template "descheduler.fullname" . }}
+subjects:
+  - kind: ServiceAccount
+    name: {{ template "descheduler.serviceAccountName" . }}
+    namespace: {{ .Release.Namespace }}
+{{- end -}}

charts/descheduler/templates/configmap.yaml

@@ -0,0 +1,11 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: {{ template "descheduler.fullname" . }}
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+data:
+  policy.yaml: |
+    apiVersion: "descheduler/v1alpha1"
+    kind: "DeschedulerPolicy"
+{{ toYaml .Values.deschedulerPolicy | trim | indent 4 }}

charts/descheduler/templates/cronjob.yaml

@@ -0,0 +1,53 @@
+apiVersion: batch/v1beta1
+kind: CronJob
+metadata:
+  name: {{ template "descheduler.fullname" . }}
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+spec:
+  schedule: {{ .Values.schedule | quote }}
+  concurrencyPolicy: "Forbid"
+  jobTemplate:
+    spec:
+      template:
+        metadata:
+          name: {{ template "descheduler.fullname" . }}
+          annotations:
+            checksum/config: {{ include (print $.Template.BasePath "/configmap.yaml") . | sha256sum }}
+            {{- if .Values.podAnnotations }}
+            {{- .Values.podAnnotations | toYaml | nindent 12 }}
+            {{- end }}
+          labels:
+            app.kubernetes.io/name: {{ include "descheduler.name" . }}
+            app.kubernetes.io/instance: {{ .Release.Name }}
+            {{- if .Values.podLabels }}
+            {{- .Values.podLabels | toYaml | nindent 12 }}
+            {{- end }}
+        spec:
+          {{- if .Values.priorityClassName }}
+          priorityClassName: {{ .Values.priorityClassName }}
+          {{- end }}
+          serviceAccountName: {{ template "descheduler.serviceAccountName" . }}
+          restartPolicy: "Never"
+          containers:
+            - name: {{ .Chart.Name }}
+              image: "{{ .Values.image.repository }}:{{ .Values.image.tag | default (printf "v%s" .Chart.AppVersion) }}"
+              imagePullPolicy: {{ .Values.image.pullPolicy }}
+              command:
+                - "/bin/descheduler"
+              args:
+                - "--policy-config-file"
+                - "/policy-dir/policy.yaml"
+                {{- range $key, $value := .Values.cmdOptions }}
+                - {{ printf "--%s" $key | quote }}
+                {{- if $value }}
+                - {{ $value | quote }}
+                {{- end }}
+                {{- end }}
+              volumeMounts:
+                - mountPath: /policy-dir
+                  name: policy-volume
+          volumes:
+          - name: policy-volume
+            configMap:
+              name: {{ template "descheduler.fullname" . }}

charts/descheduler/templates/serviceaccount.yaml

@@ -0,0 +1,8 @@
+{{- if .Values.serviceAccount.create -}}
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: {{ template "descheduler.serviceAccountName" . }}
+  labels:
+    {{- include "descheduler.labels" . | nindent 4 }}
+{{- end -}}

charts/descheduler/values.yaml

@@ -0,0 +1,59 @@
+# Default values for descheduler.
+# This is a YAML-formatted file.
+# Declare variables to be passed into your templates.
+
+image:
+  repository: us.gcr.io/k8s-artifacts-prod/descheduler/descheduler
+  # Overrides the image tag whose default is the chart version
+  tag: ""
+  pullPolicy: IfNotPresent
+
+nameOverride: ""
+fullnameOverride: ""
+
+schedule: "*/2 * * * *"
+
+cmdOptions:
+  v: 3
+  # evict-local-storage-pods:
+  # max-pods-to-evict-per-node: 10
+  # node-selector: "key1=value1,key2=value2"
+
+deschedulerPolicy:
+  strategies:
+    RemoveDuplicates:
+      enabled: true
+    RemovePodsViolatingNodeTaints:
+      enabled: true
+    RemovePodsViolatingNodeAffinity:
+      enabled: true
+      params:
+         nodeAffinityType:
+         - requiredDuringSchedulingIgnoredDuringExecution
+    RemovePodsViolatingInterPodAntiAffinity:
+      enabled: true
+    LowNodeUtilization:
+      enabled: true
+      params:
+        nodeResourceUtilizationThresholds:
+          thresholds:
+            cpu: 20
+            memory: 20
+            pods: 20
+          targetThresholds:
+            cpu: 50
+            memory: 50
+            pods: 50
+
+priorityClassName: system-cluster-critical
+
+rbac:
+  # Specifies whether RBAC resources 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:

docs/README.md

@@ -0,0 +1,5 @@
+# Documentation Index
+
+- [Contributor Guide](contributor-guide.md)
+- [Release Guide](release-guide.md)
+- [User Guide](user-guide.md)

docs/contributor-guide.md

@@ -0,0 +1,42 @@
+# Contributor Guide
+
+## Required Tools
+
+- [Git](https://git-scm.com/downloads)
+- [Go 1.13+](https://golang.org/dl/)
+- [Docker](https://docs.docker.com/install/)
+- [kubectl](https://kubernetes.io/docs/tasks/tools/install-kubectl)
+- [kind](https://kind.sigs.k8s.io/)
+
+## Build and Run
+
+Build descheduler.
+```sh
+cd $GOPATH/src/sigs.k8s.io
+git clone https://github.com/kubernetes-sigs/descheduler.git
+cd descheduler
+make
+```
+
+Run descheduler.
+```sh
+./_output/bin/descheduler --kubeconfig <path to kubeconfig> --policy-config-file <path-to-policy-file> --v 1
+```
+
+View all CLI options.
+```
+./_output/bin/descheduler --help
+```
+
+## Run Tests
+```
+GOOS=linux make dev-image
+kind create cluster --config hack/kind_config.yaml
+kind load docker-image <image name>
+kind get kubeconfig > /tmp/admin.conf
+make test-unit
+make test-e2e
+```
+
+### Miscellaneous
+See the [hack directory](https://github.com/kubernetes-sigs/descheduler/tree/master/hack) for additional tools and scripts used for developing the descheduler.

docs/release-guide.md

@@ -0,0 +1,66 @@
+# Release Guide
+
+## Container Image
+
+### Semi-automatic
+
+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 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/master/k8s.gcr.io#image-promoter)
+7. Publish release
+8. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
+
+### Manual
+
+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 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`
+7. Publish a draft release using the tag you just created
+8. Perform the [image promotion process](https://github.com/kubernetes/k8s.io/tree/master/k8s.gcr.io#image-promoter)
+9. Publish release
+10. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
+
+### Notes
+See [post-descheduler-push-images dashboard](https://testgrid.k8s.io/sig-scheduling#post-descheduler-push-images) for staging registry image build job status.
+
+List images in staging registry.
+```
+gcloud container images list --repository gcr.io/k8s-staging-descheduler
+```
+
+List descheduler image tags in the staging registry.
+```
+gcloud container images list-tags gcr.io/k8s-staging-descheduler/descheduler
+```
+
+Get SHA256 hash for a specific image in the staging registry.
+```
+gcloud container images describe gcr.io/k8s-staging-descheduler/descheduler:v20200206-0.9.0-94-ge2a23f284
+```
+
+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 `chart-*`. Example git tag name is `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 a `chart-*` git tag is created.
+
+The major and minor version of the chart matches the descheduler major and minor versions. For example descheduler helm chart version 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 appropriate release branch(i.e. release-1.18)
+   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. Create the tag and push it `git checkout release-1.18; CHART_VERSION=chart-0.18.0; git tag $CHART_VERSION; git push origin $CHART_VERSION`
+4. Verify the new helm artifact has been successfully pushed to the `gh-pages` branch

docs/releasing.md

@@ -1,21 +0,0 @@
-# Release process
-
-## Semi-automatic
-
-1. Make sure your repo is clean by git's standards
-2. Tag the repository and push the tag `VERSION=v0.10.0 git tag -m $VERSION $VERSION; git push origin $VERSION`
-3. Publish a draft release using the tag you just created
-4. Perform the [image promotion process](https://github.com/kubernetes/k8s.io/tree/master/k8s.gcr.io#image-promoter)
-5. Publish release
-6. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release
-
-## Manual
-
-1. Make sure your repo is clean by git's standards
-2. Tag the repository and push the tag `VERSION=v0.10.0 git tag -m $VERSION $VERSION; git push origin $VERSION`
-3. Checkout the tag you just created and make sure your repo is clean by git's standards `git checkout $VERSION`
-4. Build and push the container image to the staging registry `VERSION=$VERSION make push`
-5. Publish a draft release using the tag you just created
-6. Perform the [image promotion process](https://github.com/kubernetes/k8s.io/tree/master/k8s.gcr.io#image-promoter)
-7. Publish release
-8. Email `kubernetes-sig-scheduling@googlegroups.com` to announce the release

docs/user-guide.md

@@ -0,0 +1,90 @@
+# User Guide
+
+Starting with descheduler release v0.10.0 container images are available in these container registries.
+* `asia.gcr.io/k8s-artifacts-prod/descheduler/descheduler`
+* `eu.gcr.io/k8s-artifacts-prod/descheduler/descheduler`
+* `us.gcr.io/k8s-artifacts-prod/descheduler/descheduler`
+
+## Policy Configuration Examples
+The [examples](https://github.com/kubernetes-sigs/descheduler/tree/master/examples) directory has descheduler policy configuration examples.
+
+## CLI Options
+The descheduler has many CLI options that can be used to override its default behavior.
+```
+descheduler --help
+The descheduler evicts pods which may be bound to less desired nodes
+
+Usage:
+  descheduler [flags]
+  descheduler [command]
+
+Available Commands:
+  help        Help about any command
+  version     Version of descheduler
+
+Flags:
+      --add-dir-header                   If true, adds the file directory to the header
+      --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         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   Limits the maximum number of pods to be evicted per node by descheduler
+      --node-selector string             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
+
+Use "descheduler [command] --help" for more information about a command.
+```
+
+## Production Use Cases
+This section contains descriptions of real world production use cases.
+
+### Balance Cluster By Pod Age
+When initially migrating applications from a static virtual machine infrastructure to a cloud native k8s
+infrastructure there can be a tendency to treat application pods like static virtual machines. One approach
+to help prevent developers and operators from treating pods like virtual machines is to ensure that pods
+only run for a fixed amount
+of time.
+
+The `PodLifeTime` strategy can be used to ensure that old pods are evicted. It is recommended to create a
+[pod disruption budget](https://kubernetes.io/docs/tasks/run-application/configure-pdb/) for each
+application to ensure application availability.
+```
+descheduler -v=3 --evict-local-storage-pods --policy-config-file=pod-life-time.yml
+```
+
+This policy configuration file ensures that pods created more than 7 days ago are evicted.
+```
+---
+apiVersion: "descheduler/v1alpha1"
+kind: "DeschedulerPolicy"
+strategies:
+  "LowNodeUtilization":
+    enabled: false
+  "RemoveDuplicates":
+    enabled: false
+  "RemovePodsViolatingInterPodAntiAffinity":
+    enabled: false
+  "RemovePodsViolatingNodeAffinity":
+    enabled: false
+  "RemovePodsViolatingNodeTaints":
+    enabled: false
+  "RemovePodsHavingTooManyRestarts":
+    enabled: false
+  "PodLifeTime":
+    enabled: true
+    params:
+      maxPodLifeTimeSeconds: 604800 # pods run for a maximum of 7 days
+```

examples/pod-life-time.yml

@@ -0,0 +1,20 @@
+---
+apiVersion: "descheduler/v1alpha1"
+kind: "DeschedulerPolicy"
+strategies:
+  "LowNodeUtilization":
+    enabled: false
+  "RemoveDuplicates":
+    enabled: false
+  "RemovePodsViolatingInterPodAntiAffinity":
+    enabled: false
+  "RemovePodsViolatingNodeAffinity":
+    enabled: false
+  "RemovePodsViolatingNodeTaints":
+    enabled: false
+  "RemovePodsHavingTooManyRestarts":
+    enabled: false
+  "PodLifeTime":
+    enabled: true
+    params:
+      maxPodLifeTimeSeconds: 604800 # 7 days

examples/policy.yaml

@@ -17,3 +17,9 @@ strategies:
            "cpu" : 50
            "memory": 50
            "pods": 50
+  "RemovePodsHavingTooManyRestarts":
+     enabled: true
+     params:
+       podsHavingTooManyRestarts:
+         podRestartThresholds: 100
+         includingInitContainers: true

go.mod

@@ -5,10 +5,10 @@ go 1.13
 require (
 	github.com/spf13/cobra v0.0.5
 	github.com/spf13/pflag v1.0.5
-	k8s.io/api v0.17.0
-	k8s.io/apimachinery v0.17.3-beta.0
-	k8s.io/apiserver v0.17.0
-	k8s.io/client-go v0.17.0
-	k8s.io/component-base v0.17.0
+	k8s.io/api v0.18.2
+	k8s.io/apimachinery v0.18.2
+	k8s.io/apiserver v0.18.2
+	k8s.io/client-go v0.18.2
+	k8s.io/component-base v0.18.2
 	k8s.io/klog v1.0.0
 )

go.sum

@@ -29,6 +29,7 @@ github.com/beorn7/perks v0.0.0-20180321164747-3a771d992973/go.mod h1:Dwedo/Wpr24
 github.com/beorn7/perks v1.0.0/go.mod h1:KWe93zE9D1o94FZ5RNwFwVgaQK1VOXiVxmqh+CedLV8=
 github.com/bgentry/speakeasy v0.1.0/go.mod h1:+zsyZBPWlz7T6j88CTgSN5bM796AkVf0kBD4zp0CCIs=
 github.com/blang/semver v3.5.0+incompatible/go.mod h1:kRBLl5iJ+tD4TcOOxsy/0fnwebNt5EWlYSAyrTnjyyk=
+github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
 github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
 github.com/cockroachdb/datadriven v0.0.0-20190809214429-80d97fb3cbaa/go.mod h1:zn76sxSg3SzpJ0PPJaLDCu+Bu0Lg3sKTORVIj19EIF8=
 github.com/coreos/etcd v3.3.10+incompatible/go.mod h1:uF7uidLiAD3TWHmW31ZFd/JWoc32PjwdhPthX9715RE=
@@ -42,7 +43,6 @@ github.com/coreos/pkg v0.0.0-20160727233714-3ac0863d7acf/go.mod h1:E3G3o1h8I7cfc
 github.com/coreos/pkg v0.0.0-20180108230652-97fdf19511ea/go.mod h1:E3G3o1h8I7cfcXa63jLwjI0eiQQMgzzUDFVpN/nH/eA=
 github.com/cpuguy83/go-md2man v1.0.10/go.mod h1:SmD6nW6nTyfqj6ABTjUi3V3JVMnlJmwcJI5acqYI6dE=
 github.com/creack/pty v1.1.7/go.mod h1:lj5s0c3V2DBrqTV7llrYr5NG6My20zk30Fl46Y7DoTY=
-github.com/davecgh/go-spew v0.0.0-20151105211317-5215b55f46b2/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
@@ -52,9 +52,11 @@ github.com/docker/docker v0.7.3-0.20190327010347-be7ac8be2ae0/go.mod h1:eEKB0N0r
 github.com/docker/spdystream v0.0.0-20160310174837-449fdfce4d96/go.mod h1:Qh8CwZgvJUkLughtfhJv5dyTYa91l1fOUCrgjqmcifM=
 github.com/dustin/go-humanize v0.0.0-20171111073723-bb3d318650d4/go.mod h1:HtrtbFcZ19U5GC7JDqmcUSB87Iq5E25KnS6fMYU6eOk=
 github.com/dustin/go-humanize v1.0.0/go.mod h1:HtrtbFcZ19U5GC7JDqmcUSB87Iq5E25KnS6fMYU6eOk=
-github.com/elazarl/goproxy v0.0.0-20170405201442-c4fc26588b6e/go.mod h1:/Zj4wYkgs4iZTTu3o/KG3Itv/qCCa8VVMlb3i9OVuzc=
+github.com/elazarl/goproxy v0.0.0-20180725130230-947c36da3153/go.mod h1:/Zj4wYkgs4iZTTu3o/KG3Itv/qCCa8VVMlb3i9OVuzc=
 github.com/emicklei/go-restful v0.0.0-20170410110728-ff4f55a20633/go.mod h1:otzb+WCGbkyDHkqmQmT5YD2WR4BBwUdeQoFo8l/7tVs=
 github.com/emicklei/go-restful v2.9.5+incompatible/go.mod h1:otzb+WCGbkyDHkqmQmT5YD2WR4BBwUdeQoFo8l/7tVs=
+github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
 github.com/evanphx/json-patch v4.2.0+incompatible h1:fUDGZCv/7iAN7u0puUVhvKCcsR6vRfwrJatElLBEf0I=
 github.com/evanphx/json-patch v4.2.0+incompatible/go.mod h1:50XU6AFN0ol/bzJsmQLiYLvXMP4fmwYFNcr97nuDLSk=
 github.com/fatih/color v1.7.0/go.mod h1:Zm6kSWBoL9eyXnKyktHP6abPY2pDugNf5KwzbycvMj4=
@@ -79,8 +81,8 @@ github.com/go-openapi/swag v0.19.5/go.mod h1:POnQmlKehdgb5mhVOsnJFsivZCEZ/vjK9gh
 github.com/go-stack/stack v1.8.0/go.mod h1:v0f6uXyyMGvRgIKkXu+yp6POWl0qKG85gN/melR3HDY=
 github.com/gogo/protobuf v1.1.1/go.mod h1:r8qH/GZQm5c6nD/R0oafs1akxWv10x8SbQlK7atdtwQ=
 github.com/gogo/protobuf v1.2.1/go.mod h1:hp+jE20tsWTFYpLwKvXlhS1hjn+gTNwPg2I6zVXpSg4=
-github.com/gogo/protobuf v1.2.2-0.20190723190241-65acae22fc9d h1:3PaI8p3seN09VjbTYC/QWlUZdZ1qS1zGjy7LH2Wt07I=
-github.com/gogo/protobuf v1.2.2-0.20190723190241-65acae22fc9d/go.mod h1:SlYgWuQ5SjCEi6WLHjHCa1yvBfUnHcTbrrZtXPKa29o=
+github.com/gogo/protobuf v1.3.1 h1:DqDEcV5aeaTmdFBePNpYsp3FlcVH/2ISVVM9Qf8PSls=
+github.com/gogo/protobuf v1.3.1/go.mod h1:SlYgWuQ5SjCEi6WLHjHCa1yvBfUnHcTbrrZtXPKa29o=
 github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
 github.com/golang/groupcache v0.0.0-20160516000752-02826c3e7903 h1:LbsanbbD6LieFkXbj9YNNBupiGHJgFeLpO0j0Fza1h8=
 github.com/golang/groupcache v0.0.0-20160516000752-02826c3e7903/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
@@ -96,9 +98,10 @@ github.com/google/btree v1.0.0/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ
 github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
 github.com/google/go-cmp v0.3.0 h1:crn/baboCvb5fXaQ0IJ1SGTsTVrWpDsCWC8EGETZijY=
 github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
-github.com/google/gofuzz v0.0.0-20161122191042-44d81051d367/go.mod h1:HP5RmnzzSNb993RKQDq4+1A4ia9nllfqcQFTQJedwGI=
 github.com/google/gofuzz v1.0.0 h1:A8PeW59pxE9IoFRqBp37U+mSNaQoZ46F1f0f863XSXw=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
+github.com/google/gofuzz v1.1.0 h1:Hsa8mG0dQ46ij8Sl2AYJDUv1oA9/d6Vk+3LG99Oe02g=
+github.com/google/gofuzz v1.1.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
 github.com/google/martian v2.1.0+incompatible/go.mod h1:9I4somxYTbIHy5NJKHRl3wXiIaQGbYVAs8BPL6v8lEs=
 github.com/google/pprof v0.0.0-20181206194817-3ea8567a2e57/go.mod h1:zfwlbNMJ+OItoe0UupaVj+oy1omPYYDuagoSzA8v9mc=
 github.com/google/uuid v1.0.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
@@ -107,6 +110,8 @@ github.com/google/uuid v1.1.1/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+
 github.com/googleapis/gax-go/v2 v2.0.4/go.mod h1:0Wqv26UfaUD9n4G6kQubkQ+KchISgw+vpHVxEJEs9eg=
 github.com/googleapis/gnostic v0.0.0-20170729233727-0c5108395e2d h1:7XGaL1e6bYS1yIonGp9761ExpPPV1ui0SAC59Yube9k=
 github.com/googleapis/gnostic v0.0.0-20170729233727-0c5108395e2d/go.mod h1:sJBsCZ4ayReDTBIg8b9dl28c5xFWyhBTVRp3pOg5EKY=
+github.com/googleapis/gnostic v0.1.0 h1:rVsPeBmXbYv4If/cumu1AzZPwV58q433hvONV1UEZoI=
+github.com/googleapis/gnostic v0.1.0/go.mod h1:sJBsCZ4ayReDTBIg8b9dl28c5xFWyhBTVRp3pOg5EKY=
 github.com/gophercloud/gophercloud v0.1.0 h1:P/nh25+rzXouhytV2pUHBb65fnds26Ghl8/391+sT5o=
 github.com/gophercloud/gophercloud v0.1.0/go.mod h1:vxM41WHh5uqHVBMZHzuwNOHh8XEoIEcSTewFxm1c5g8=
 github.com/gorilla/websocket v0.0.0-20170926233335-4201258b820c/go.mod h1:E7qHFY5m1UJ88s3WnNqhKjPHQ0heANvMoAMk2YaljkQ=
@@ -126,7 +131,6 @@ github.com/imdario/mergo v0.3.5/go.mod h1:2EnlNZ0deacrJVfApfmtdGgDfMuh/nq6Ok1EcJ
 github.com/inconshreveable/mousetrap v1.0.0 h1:Z8tu5sraLXCXIcARxBp/8cbvlwVa7Z1NHg9XEKhtSvM=
 github.com/inconshreveable/mousetrap v1.0.0/go.mod h1:PxqpIevigyE2G7u3NXJIT2ANytuPF1OarO4DADm73n8=
 github.com/jonboulle/clockwork v0.1.0/go.mod h1:Ii8DK3G1RaLaWxj9trq07+26W01tbo22gdxWY5EU2bo=
-github.com/json-iterator/go v0.0.0-20180612202835-f2b4162afba3/go.mod h1:+SdeFBvtyEkXs7REEP0seUULqWtbJapLOCVDaaPEHmU=
 github.com/json-iterator/go v1.1.6/go.mod h1:+SdeFBvtyEkXs7REEP0seUULqWtbJapLOCVDaaPEHmU=
 github.com/json-iterator/go v1.1.7/go.mod h1:KdQUCv79m/52Kvf8AW2vK1V8akMuk1QjK/uOdHXbAo4=
 github.com/json-iterator/go v1.1.8 h1:QiWkFLKq0T7mpzwOTu6BzNDbfTE8OLrYhVKYMLF46Ok=
@@ -158,7 +162,6 @@ github.com/mitchellh/mapstructure v1.1.2/go.mod h1:FVVH3fgwuzCH5S8UJGiWEs2h04kUh
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd h1:TRLaZ9cD/w8PVh93nsPXa1VrQ6jlwL5oN8l14QlcNfg=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
-github.com/modern-go/reflect2 v0.0.0-20180320133207-05fbef0ca5da/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
 github.com/modern-go/reflect2 v0.0.0-20180701023420-4b7aa43c6742/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
 github.com/modern-go/reflect2 v1.0.1 h1:9f412s+6RmYXLWZSEzVVgPGK7C2PphHj5RJrvfx9AWI=
 github.com/modern-go/reflect2 v1.0.1/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
@@ -169,8 +172,8 @@ github.com/mxk/go-flowrate v0.0.0-20140419014527-cca7078d478f/go.mod h1:ZdcZmHo+
 github.com/olekukonko/tablewriter v0.0.0-20170122224234-a0225b3f23b5/go.mod h1:vsDQFd/mU46D+Z4whnwzcISnGGzXWMclvtLoiIKAKIo=
 github.com/onsi/ginkgo v0.0.0-20170829012221-11459a886d9c/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
 github.com/onsi/ginkgo v1.6.0/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
-github.com/onsi/ginkgo v1.10.1 h1:q/mM8GF/n0shIN8SaAZ0V+jnLPzen6WIVZdiwrRlMlo=
-github.com/onsi/ginkgo v1.10.1/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
+github.com/onsi/ginkgo v1.11.0 h1:JAKSXpt1YjtLA7YpPiqO9ss6sNXEsPfSGdwN0UHqzrw=
+github.com/onsi/ginkgo v1.11.0/go.mod h1:lLunBs/Ym6LB5Z9jYTR76FiuTmxDTDusOGeTQH+WWjE=
 github.com/onsi/gomega v0.0.0-20170829124025-dcabb60a477c/go.mod h1:C1qb7wdrVGGVU+Z6iS04AVkA3Q65CEZX59MT0QO5uiA=
 github.com/onsi/gomega v1.7.0 h1:XPnZz8VVBHjVsy1vzJmRwIcSwiUO+JFfrv/xGiigmME=
 github.com/onsi/gomega v1.7.0/go.mod h1:ex+gbHU/CVuBBDIJjb2X0qEXbFg53c61hWP/1CpauHY=
@@ -178,7 +181,6 @@ github.com/pelletier/go-toml v1.2.0/go.mod h1:5z9KED0ma1S8pY6P1sdut58dfprrGBbd/9
 github.com/peterbourgon/diskv v2.0.1+incompatible/go.mod h1:uqqh8zWWbv1HBMNONnaR/tNboyR3/BZd58JJSHlUSCU=
 github.com/pkg/errors v0.8.0/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pkg/errors v0.8.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
-github.com/pmezard/go-difflib v0.0.0-20151028094244-d8ed2627bdf0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/pquerna/cachecontrol v0.0.0-20171018203845-0dec1b30a021/go.mod h1:prYjPmNq4d1NPVmpShWobRqXY3q7Vp+80DqgxxUrUIA=
@@ -186,6 +188,8 @@ github.com/prometheus/client_golang v0.9.1/go.mod h1:7SWBe2y4D6OKWSNQJUaRYU/AaXP
 github.com/prometheus/client_golang v1.0.0/go.mod h1:db9x61etRT2tGnBNRi70OPL5FsnadC4Ky3P0J6CfImo=
 github.com/prometheus/client_model v0.0.0-20180712105110-5c3871d89910/go.mod h1:MbSGuTsp3dbXC40dX6PRTWyKYBIrTGTE9sqQNg2J8bo=
 github.com/prometheus/client_model v0.0.0-20190129233127-fd36f4220a90/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
+github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
+github.com/prometheus/client_model v0.2.0/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
 github.com/prometheus/common v0.4.1/go.mod h1:TNfzLD0ON7rHzMJeJkieUDPYmFC7Snx/y86RQel1bk4=
 github.com/prometheus/procfs v0.0.0-20181005140218-185b4288413d/go.mod h1:c3At6R/oaqEKCNdg8wHV1ftS6bRYblBhIjjI8uT2IGk=
 github.com/prometheus/procfs v0.0.2/go.mod h1:TjEm7ze935MbeOT/UhFTIMYKhuLP4wbCsTZCD3I8kEA=
@@ -210,7 +214,6 @@ github.com/spf13/viper v1.3.2/go.mod h1:ZiWeW+zYFKm7srdB9IoDzzZXaJaI5eL9QjNiN/DM
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/objx v0.1.1/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/objx v0.2.0/go.mod h1:qt09Ya8vawLte6SNmTgCsAVtYtaKzEcn8ATUoHMkEqE=
-github.com/stretchr/testify v0.0.0-20151208002404-e3a8ff8ce365/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
 github.com/stretchr/testify v1.4.0 h1:2E4SXV/wtOkTonXsotYi4li6zVWxYlZuYNCXe9XRJyk=
@@ -231,8 +234,8 @@ golang.org/x/crypto v0.0.0-20181203042331-505ab145d0a9/go.mod h1:6SG95UA2DQfeDnf
 golang.org/x/crypto v0.0.0-20190211182817-74369b46fc67/go.mod h1:6SG95UA2DQfeDnfUPMdvaQW0Q7yPrPDi9nlGo2tz2b4=
 golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
 golang.org/x/crypto v0.0.0-20190611184440-5c40567a22f8/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
-golang.org/x/crypto v0.0.0-20190820162420-60c769a6c586 h1:7KByu05hhLed2MO29w7p1XfZvZ13m8mub3shuVftRs0=
-golang.org/x/crypto v0.0.0-20190820162420-60c769a6c586/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20200220183623-bac4c82f6975 h1:/Tl7pH94bvbAAHBdZJT947M/+gp0+CqQXDtMRC0fseo=
+golang.org/x/crypto v0.0.0-20200220183623-bac4c82f6975/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
 golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
 golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
 golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
@@ -262,6 +265,7 @@ golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJ
 golang.org/x/sync v0.0.0-20181221193216-37e7f081c4d4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20190227155943-e225da77a7e6/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190911185100-cd5d95a43a6e/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sys v0.0.0-20170830134202-bb24a47a89ea/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
@@ -276,6 +280,8 @@ golang.org/x/sys v0.0.0-20190422165155-953cdadca894/go.mod h1:h1NjWce9XRLGQEsW7w
 golang.org/x/sys v0.0.0-20190616124812-15dcb6c0061f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20190826190057-c7b8b68b1456 h1:ng0gs1AKnRRuEMZoTLLlbOd+C17zUDepwGQBb/n+JVg=
 golang.org/x/sys v0.0.0-20190826190057-c7b8b68b1456/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191022100944-742c48ecaeb7 h1:HmbHVPwrPEKPGLAcHSrMe6+hqSUlvZU0rab6x5EXfGU=
+golang.org/x/sys v0.0.0-20191022100944-742c48ecaeb7/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/text v0.0.0-20160726164857-2910a502d2bf/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.1-0.20180807135948-17ff2d5776d2/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
@@ -303,9 +309,11 @@ google.golang.org/appengine v1.5.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7
 google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
 google.golang.org/genproto v0.0.0-20190307195333-5fe7a883aa19/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
 google.golang.org/genproto v0.0.0-20190418145605-e7d98fc518a7/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
-google.golang.org/genproto v0.0.0-20190502173448-54afdca5d873/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
 google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
+google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
 google.golang.org/grpc v1.23.1/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
+google.golang.org/grpc v1.26.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
 gopkg.in/alecthomas/kingpin.v2 v2.2.6/go.mod h1:FMv+mEhP44yOT+4EoQTLFTRgOQ1FBLkstjWtayDeSgw=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 h1:qIbj1fsPNlZgppZ+VLlY7N33q108Sa+fhmuc+sWQYwY=
@@ -323,33 +331,36 @@ gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7/go.mod h1:dt/ZhP58zS4L8KSrWD
 gopkg.in/yaml.v2 v2.0.0-20170812160011-eb3733d160e7/go.mod h1:JAlM8MvJe8wmxCU4Bli9HhUf9+ttbYbLASfIpnQbh74=
 gopkg.in/yaml.v2 v2.2.1/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
-gopkg.in/yaml.v2 v2.2.4 h1:/eiJrUcujPVeJ3xlSWaiNi3uSVmDGBK1pDHUHAnao1I=
-gopkg.in/yaml.v2 v2.2.4/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v2 v2.2.8 h1:obN1ZagJSUGI0Ek/LBmuj4SNLPfIny3KsKFopxRdj10=
+gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gotest.tools v2.2.0+incompatible/go.mod h1:DsYFclhRJ6vuDpmuTbkuFWG+y2sxOXAzmJt81HFBacw=
 honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
 honnef.co/go/tools v0.0.0-20190106161140-3f1c8253044a/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
 honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
-k8s.io/api v0.17.0 h1:H9d/lw+VkZKEVIUc8F3wgiQ+FUXTTr21M87jXLU7yqM=
-k8s.io/api v0.17.0/go.mod h1:npsyOePkeP0CPwyGfXDHxvypiYMJxBWAMpQxCaJ4ZxI=
-k8s.io/apimachinery v0.17.0/go.mod h1:b9qmWdKlLuU9EBh+06BtLcSf/Mu89rWL33naRxs1uZg=
-k8s.io/apimachinery v0.17.3-beta.0 h1:DeN0royOQ5+j3ytFWnDkxGiF5r9T7m6E9Ukzjg4vVHc=
-k8s.io/apimachinery v0.17.3-beta.0/go.mod h1:b9qmWdKlLuU9EBh+06BtLcSf/Mu89rWL33naRxs1uZg=
-k8s.io/apiserver v0.17.0 h1:XhUix+FKFDcBygWkQNp7wKKvZL030QUlH1o8vFeSgZA=
-k8s.io/apiserver v0.17.0/go.mod h1:ABM+9x/prjINN6iiffRVNCBR2Wk7uY4z+EtEGZD48cg=
-k8s.io/client-go v0.17.0 h1:8QOGvUGdqDMFrm9sD6IUFl256BcffynGoe80sxgTEDg=
-k8s.io/client-go v0.17.0/go.mod h1:TYgR6EUHs6k45hb6KWjVD6jFZvJV4gHDikv/It0xz+k=
-k8s.io/component-base v0.17.0 h1:BnDFcmBDq+RPpxXjmuYnZXb59XNN9CaFrX8ba9+3xrA=
-k8s.io/component-base v0.17.0/go.mod h1:rKuRAokNMY2nn2A6LP/MiwpoaMRHpfRnrPaUJJj1Yoc=
+k8s.io/api v0.18.2 h1:wG5g5ZmSVgm5B+eHMIbI9EGATS2L8Z72rda19RIEgY8=
+k8s.io/api v0.18.2/go.mod h1:SJCWI7OLzhZSvbY7U8zwNl9UA4o1fizoug34OV/2r78=
+k8s.io/apimachinery v0.18.2 h1:44CmtbmkzVDAhCpRVSiP2R5PPrC2RtlIv/MoB8xpdRA=
+k8s.io/apimachinery v0.18.2/go.mod h1:9SnR/e11v5IbyPCGbvJViimtJ0SwHG4nfZFjU77ftcA=
+k8s.io/apiserver v0.18.2 h1:fwKxdTWwwYhxvtjo0UUfX+/fsitsNtfErPNegH2x9ic=
+k8s.io/apiserver v0.18.2/go.mod h1:Xbh066NqrZO8cbsoenCwyDJ1OSi8Ag8I2lezeHxzwzw=
+k8s.io/client-go v0.18.2 h1:aLB0iaD4nmwh7arT2wIn+lMnAq7OswjaejkQ8p9bBYE=
+k8s.io/client-go v0.18.2/go.mod h1:Xcm5wVGXX9HAA2JJ2sSBUn3tCJ+4SVlCbl2MNNv+CIU=
+k8s.io/component-base v0.18.2 h1:SJweNZAGcUvsypLGNPNGeJ9UgPZQ6+bW+gEHe8uyh/Y=
+k8s.io/component-base v0.18.2/go.mod h1:kqLlMuhJNHQ9lz8Z7V5bxUUtjFZnrypArGl58gmDfUM=
 k8s.io/gengo v0.0.0-20190128074634-0689ccc1d7d6/go.mod h1:ezvh/TsK7cY6rbqRK0oQQ8IAqLxYwwyPxAX1Pzy0ii0=
 k8s.io/klog v0.0.0-20181102134211-b9b56d5dfc92/go.mod h1:Gq+BEi5rUBO/HRz0bTSXDUcqjScdoY3a9IHpCEIOOfk=
 k8s.io/klog v0.3.0/go.mod h1:Gq+BEi5rUBO/HRz0bTSXDUcqjScdoY3a9IHpCEIOOfk=
 k8s.io/klog v1.0.0 h1:Pt+yjF5aB1xDSVbau4VsWe+dQNzA0qv1LlXdC2dF6Q8=
 k8s.io/klog v1.0.0/go.mod h1:4Bi6QPql/J/LkTDqv7R/cd3hPo4k2DG6Ptcz060Ez5I=
-k8s.io/kube-openapi v0.0.0-20191107075043-30be4d16710a h1:UcxjrRMyNx/i/y8G7kPvLyy7rfbeuf1PYyBf973pgyU=
-k8s.io/kube-openapi v0.0.0-20191107075043-30be4d16710a/go.mod h1:1TqjTSzOxsLGIKfj0lK8EeCP7K1iUG65v09OM0/WG5E=
-k8s.io/utils v0.0.0-20191114184206-e782cd3c129f h1:GiPwtSzdP43eI1hpPCbROQCCIgCuiMMNF8YUVLF3vJo=
-k8s.io/utils v0.0.0-20191114184206-e782cd3c129f/go.mod h1:sZAwmy6armz5eXlNoLmJcl4F1QuKu7sr+mFQ0byX7Ew=
-sigs.k8s.io/structured-merge-diff v0.0.0-20190525122527-15d366b2352e/go.mod h1:wWxsB5ozmmv/SG7nM11ayaAW51xMvak/t1r0CSlcokI=
-sigs.k8s.io/structured-merge-diff v1.0.1-0.20191108220359-b1b620dd3f06/go.mod h1:/ULNhyfzRopfcjskuui0cTITekDduZ7ycKN3oUT9R18=
+k8s.io/kube-openapi v0.0.0-20200121204235-bf4fb3bd569c h1:/KUFqjjqAcY4Us6luF5RDNZ16KJtb49HfR3ZHB9qYXM=
+k8s.io/kube-openapi v0.0.0-20200121204235-bf4fb3bd569c/go.mod h1:GRQhZsXIAJ1xR0C9bd8UpWHZ5plfAS9fzPjJuQ6JL3E=
+k8s.io/utils v0.0.0-20200324210504-a9aa75ae1b89 h1:d4vVOjXm687F1iLSP2q3lyPPuyvTUt3aVoBpi2DqRsU=
+k8s.io/utils v0.0.0-20200324210504-a9aa75ae1b89/go.mod h1:sZAwmy6armz5eXlNoLmJcl4F1QuKu7sr+mFQ0byX7Ew=
+sigs.k8s.io/apiserver-network-proxy/konnectivity-client v0.0.7/go.mod h1:PHgbrJT7lCHcxMU+mDHEm+nx46H4zuuHZkDP6icnhu0=
+sigs.k8s.io/structured-merge-diff/v3 v3.0.0-20200116222232-67a7b8c61874/go.mod h1:PlARxl6Hbt/+BC80dRLi1qAmnMqwqDg62YvvVkZjemw=
+sigs.k8s.io/structured-merge-diff/v3 v3.0.0 h1:dOmIZBMfhcHS09XZkMyUgkq5trg3/jRyJYFZUiaOp8E=
+sigs.k8s.io/structured-merge-diff/v3 v3.0.0/go.mod h1:PlARxl6Hbt/+BC80dRLi1qAmnMqwqDg62YvvVkZjemw=
 sigs.k8s.io/yaml v1.1.0 h1:4A07+ZFc2wgJwo8YNlQpr1rVlgUDlxXHhPJciaPY5gs=
 sigs.k8s.io/yaml v1.1.0/go.mod h1:UJmg0vDUVViEyp3mgSv9WPwZCDxu4rQW1olrI1uml+o=
+sigs.k8s.io/yaml v1.2.0 h1:kr/MCeFWJWTwyaHoR9c8EjH9OumOmoF9YGiZd7lFm/Q=
+sigs.k8s.io/yaml v1.2.0/go.mod h1:yfXDCHCao9+ENCvLSE62v9VSji2MKu5jeNfTrofGhJc=

hack/update-vendor.sh

@@ -0,0 +1,21 @@
+#!/bin/bash
+
+# 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.
+
+
+source "$(dirname "${BASH_SOURCE}")/lib/init.sh"
+
+go mod tidy
+go mod vendor

hack/verify-vendor.sh

@@ -0,0 +1,88 @@
+#!/bin/bash
+
+# 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.
+
+# This is mostly copied from the hack/verify-vendor.sh script located in k8s.io/kubernetes
+
+set -o errexit
+set -o nounset
+set -o pipefail
+
+source "$(dirname "${BASH_SOURCE}")/lib/init.sh"
+
+DESCHEDULER_ROOT=$(dirname "${BASH_SOURCE}")/..
+
+mkdir -p "${DESCHEDULER_ROOT}/_tmp"
+_tmpdir="$(mktemp -d "${DESCHEDULER_ROOT}/_tmp/kube-vendor.XXXXXX")"
+
+if [[ -z ${KEEP_TMP:-} ]]; then
+    KEEP_TMP=false
+fi
+
+function cleanup {
+  # make go module dirs writeable
+  chmod -R +w "${_tmpdir}"
+  if [ "${KEEP_TMP}" == "true" ]; then
+    echo "Leaving ${_tmpdir} for you to examine or copy. Please delete it manually when finished. (rm -rf ${_tmpdir})"
+  else
+    echo "Removing ${_tmpdir}"
+    rm -rf "${_tmpdir}"
+  fi
+}
+trap "cleanup" EXIT
+
+_deschedulertmp="${_tmpdir}"
+mkdir -p "${_deschedulertmp}"
+
+git archive --format=tar --prefix=descheduler/ "$(git write-tree)" | (cd "${_deschedulertmp}" && tar xf -)
+_deschedulertmp="${_deschedulertmp}/descheduler"
+
+pushd "${_deschedulertmp}" > /dev/null 2>&1
+  # Destroy deps in the copy of the kube tree
+  rm -rf ./vendor
+
+  # Recreate the vendor tree using the nice clean set we just downloaded
+  hack/update-vendor.sh
+popd > /dev/null 2>&1
+
+ret=0
+
+pushd "${DESCHEDULER_ROOT}" > /dev/null 2>&1
+  # Test for diffs
+  if ! _out="$(diff -Naupr --ignore-matching-lines='^\s*\"GoVersion\":' go.mod "${_deschedulertmp}/go.mod")"; then
+    echo "Your go.mod file is different:" >&2
+    echo "${_out}" >&2
+    echo "Vendor Verify failed." >&2
+    echo "If you're seeing this locally, run the below command to fix your go.mod:" >&2
+    echo "hack/update-vendor.sh" >&2
+    ret=1
+  fi
+
+  if ! _out="$(diff -Naupr -x "BUILD" -x "AUTHORS*" -x "CONTRIBUTORS*" vendor "${_deschedulertmp}/vendor")"; then
+    echo "Your vendored results are different:" >&2
+    echo "${_out}" >&2
+    echo "Vendor Verify failed." >&2
+    echo "${_out}" > vendordiff.patch
+    echo "If you're seeing this locally, run the below command to fix your directories:" >&2
+    echo "hack/update-vendor.sh" >&2
+    ret=1
+  fi
+popd > /dev/null 2>&1
+
+if [[ ${ret} -gt 0 ]]; then
+  exit ${ret}
+fi
+
+echo "Vendor Verified."

kubernetes/cronjob.yaml

@@ -16,7 +16,7 @@ spec:
           priorityClassName: system-cluster-critical
           containers:
           - name: descheduler
-            image: us.gcr.io/k8s-artifacts-prod/descheduler:v0.10.0
+            image: us.gcr.io/k8s-artifacts-prod/descheduler/descheduler:v0.18.0
             volumeMounts:
             - mountPath: /policy-dir
               name: policy-volume

kubernetes/job.yaml

@@ -14,7 +14,7 @@ spec:
       priorityClassName: system-cluster-critical
       containers:
         - name: descheduler
-          image: us.gcr.io/k8s-artifacts-prod/descheduler:v0.10.0
+          image: us.gcr.io/k8s-artifacts-prod/descheduler/descheduler:v0.18.0
           volumeMounts:
           - mountPath: /policy-dir
             name: policy-volume

pkg/api/types.go

@@ -17,7 +17,7 @@ limitations under the License.
 package api
 
 import (
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 )
 
@@ -46,8 +46,11 @@ type DeschedulerStrategy struct {
 
 // Only one of its members may be specified
 type StrategyParameters struct {
-	NodeResourceUtilizationThresholds NodeResourceUtilizationThresholds
+	NodeResourceUtilizationThresholds *NodeResourceUtilizationThresholds
 	NodeAffinityType                  []string
+	PodsHavingTooManyRestarts         *PodsHavingTooManyRestarts
+	MaxPodLifeTimeSeconds             *uint
+	RemoveDuplicates                  *RemoveDuplicates
 }
 
 type Percentage float64
@@ -58,3 +61,12 @@ type NodeResourceUtilizationThresholds struct {
 	TargetThresholds ResourceThresholds
 	NumberOfNodes    int
 }
+
+type PodsHavingTooManyRestarts struct {
+	PodRestartThreshold     int32
+	IncludingInitContainers bool
+}
+
+type RemoveDuplicates struct {
+	ExcludeOwnerKinds []string
+}

pkg/api/v1alpha1/types.go

@@ -17,7 +17,7 @@ limitations under the License.
 package v1alpha1
 
 import (
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 )
 
@@ -46,8 +46,11 @@ type DeschedulerStrategy struct {
 
 // Only one of its members may be specified
 type StrategyParameters struct {
-	NodeResourceUtilizationThresholds NodeResourceUtilizationThresholds `json:"nodeResourceUtilizationThresholds,omitempty"`
-	NodeAffinityType                  []string                          `json:"nodeAffinityType,omitempty"`
+	NodeResourceUtilizationThresholds *NodeResourceUtilizationThresholds `json:"nodeResourceUtilizationThresholds,omitempty"`
+	NodeAffinityType                  []string                           `json:"nodeAffinityType,omitempty"`
+	PodsHavingTooManyRestarts         *PodsHavingTooManyRestarts         `json:"podsHavingTooManyRestarts,omitempty"`
+	MaxPodLifeTimeSeconds             *uint                              `json:"maxPodLifeTimeSeconds,omitempty"`
+	RemoveDuplicates                  *RemoveDuplicates                  `json:"removeDuplicates,omitempty"`
 }
 
 type Percentage float64
@@ -58,3 +61,12 @@ type NodeResourceUtilizationThresholds struct {
 	TargetThresholds ResourceThresholds `json:"targetThresholds,omitempty"`
 	NumberOfNodes    int                `json:"numberOfNodes,omitempty"`
 }
+
+type PodsHavingTooManyRestarts struct {
+	PodRestartThreshold     int32 `json:"podRestartThreshold,omitempty"`
+	IncludingInitContainers bool  `json:"includingInitContainers,omitempty"`
+}
+
+type RemoveDuplicates struct {
+	ExcludeOwnerKinds []string `json:"excludeOwnerKinds,omitempty"`
+}

pkg/api/v1alpha1/zz_generated.conversion.go

@@ -65,6 +65,26 @@ func RegisterConversions(s *runtime.Scheme) error {
 	}); 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 {
@@ -150,11 +170,54 @@ func Convert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtili
 	return autoConvert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds(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 {
-	if err := Convert_v1alpha1_NodeResourceUtilizationThresholds_To_api_NodeResourceUtilizationThresholds(&in.NodeResourceUtilizationThresholds, &out.NodeResourceUtilizationThresholds, s); err != nil {
-		return err
-	}
+	out.NodeResourceUtilizationThresholds = (*api.NodeResourceUtilizationThresholds)(unsafe.Pointer(in.NodeResourceUtilizationThresholds))
 	out.NodeAffinityType = *(*[]string)(unsafe.Pointer(&in.NodeAffinityType))
+	out.PodsHavingTooManyRestarts = (*api.PodsHavingTooManyRestarts)(unsafe.Pointer(in.PodsHavingTooManyRestarts))
+	out.MaxPodLifeTimeSeconds = (*uint)(unsafe.Pointer(in.MaxPodLifeTimeSeconds))
+	out.RemoveDuplicates = (*api.RemoveDuplicates)(unsafe.Pointer(in.RemoveDuplicates))
 	return nil
 }
 
@@ -164,10 +227,11 @@ func Convert_v1alpha1_StrategyParameters_To_api_StrategyParameters(in *StrategyP
 }
 
 func autoConvert_api_StrategyParameters_To_v1alpha1_StrategyParameters(in *api.StrategyParameters, out *StrategyParameters, s conversion.Scope) error {
-	if err := Convert_api_NodeResourceUtilizationThresholds_To_v1alpha1_NodeResourceUtilizationThresholds(&in.NodeResourceUtilizationThresholds, &out.NodeResourceUtilizationThresholds, s); err != nil {
-		return err
-	}
+	out.NodeResourceUtilizationThresholds = (*NodeResourceUtilizationThresholds)(unsafe.Pointer(in.NodeResourceUtilizationThresholds))
 	out.NodeAffinityType = *(*[]string)(unsafe.Pointer(&in.NodeAffinityType))
+	out.PodsHavingTooManyRestarts = (*PodsHavingTooManyRestarts)(unsafe.Pointer(in.PodsHavingTooManyRestarts))
+	out.MaxPodLifeTimeSeconds = (*uint)(unsafe.Pointer(in.MaxPodLifeTimeSeconds))
+	out.RemoveDuplicates = (*RemoveDuplicates)(unsafe.Pointer(in.RemoveDuplicates))
 	return nil
 }
 

pkg/api/v1alpha1/zz_generated.deepcopy.go

@@ -103,6 +103,43 @@ func (in *NodeResourceUtilizationThresholds) DeepCopy() *NodeResourceUtilization
 	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) {
+	*out = *in
+	return
+}
+
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PodsHavingTooManyRestarts.
+func (in *PodsHavingTooManyRestarts) DeepCopy() *PodsHavingTooManyRestarts {
+	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)
+	in.DeepCopyInto(out)
+	return out
+}
+
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
 func (in ResourceThresholds) DeepCopyInto(out *ResourceThresholds) {
 	{
@@ -150,12 +187,31 @@ func (in StrategyList) DeepCopy() StrategyList {
 // 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
-	in.NodeResourceUtilizationThresholds.DeepCopyInto(&out.NodeResourceUtilizationThresholds)
+	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.MaxPodLifeTimeSeconds != nil {
+		in, out := &in.MaxPodLifeTimeSeconds, &out.MaxPodLifeTimeSeconds
+		*out = new(uint)
+		**out = **in
+	}
+	if in.RemoveDuplicates != nil {
+		in, out := &in.RemoveDuplicates, &out.RemoveDuplicates
+		*out = new(RemoveDuplicates)
+		(*in).DeepCopyInto(*out)
+	}
 	return
 }
 

pkg/api/zz_generated.deepcopy.go

@@ -103,6 +103,43 @@ func (in *NodeResourceUtilizationThresholds) DeepCopy() *NodeResourceUtilization
 	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) {
+	*out = *in
+	return
+}
+
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new PodsHavingTooManyRestarts.
+func (in *PodsHavingTooManyRestarts) DeepCopy() *PodsHavingTooManyRestarts {
+	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)
+	in.DeepCopyInto(out)
+	return out
+}
+
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
 func (in ResourceThresholds) DeepCopyInto(out *ResourceThresholds) {
 	{
@@ -150,12 +187,31 @@ func (in StrategyList) DeepCopy() StrategyList {
 // 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
-	in.NodeResourceUtilizationThresholds.DeepCopyInto(&out.NodeResourceUtilizationThresholds)
+	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.MaxPodLifeTimeSeconds != nil {
+		in, out := &in.MaxPodLifeTimeSeconds, &out.MaxPodLifeTimeSeconds
+		*out = new(uint)
+		**out = **in
+	}
+	if in.RemoveDuplicates != nil {
+		in, out := &in.RemoveDuplicates, &out.RemoveDuplicates
+		*out = new(RemoveDuplicates)
+		(*in).DeepCopyInto(*out)
+	}
 	return
 }
 

pkg/descheduler/descheduler.go

@@ -17,21 +17,26 @@ limitations under the License.
 package descheduler
 
 import (
+	"context"
 	"fmt"
 
+	"k8s.io/api/core/v1"
+	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog"
 
 	"k8s.io/apimachinery/pkg/util/wait"
+	"k8s.io/client-go/informers"
 	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"sigs.k8s.io/descheduler/pkg/api"
 	"sigs.k8s.io/descheduler/pkg/descheduler/client"
+	"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"
-	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
 func Run(rs *options.DeschedulerServer) error {
+	ctx := context.Background()
 	rsclient, err := client.CreateClient(rs.KubeconfigFile)
 	if err != nil {
 		return err
@@ -46,33 +51,61 @@ func Run(rs *options.DeschedulerServer) error {
 		return fmt.Errorf("deschedulerPolicy is nil")
 	}
 
-	return RunDeschedulerStrategies(rs, deschedulerPolicy)
-}
-
-func RunDeschedulerStrategies(rs *options.DeschedulerServer, deschedulerPolicy *api.DeschedulerPolicy) error {
 	evictionPolicyGroupVersion, err := eutils.SupportEviction(rs.Client)
 	if err != nil || len(evictionPolicyGroupVersion) == 0 {
 		return err
 	}
 
 	stopChannel := make(chan struct{})
-	nodes, err := nodeutil.ReadyNodes(rs.Client, rs.NodeSelector, stopChannel)
-	if err != nil {
-		return err
+	return RunDeschedulerStrategies(ctx, rs, deschedulerPolicy, evictionPolicyGroupVersion, stopChannel)
+}
+
+type strategyFunction func(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, evictLocalStoragePods bool, 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[string]strategyFunction{
+		"RemoveDuplicates":                        strategies.RemoveDuplicatePods,
+		"LowNodeUtilization":                      strategies.LowNodeUtilization,
+		"RemovePodsViolatingInterPodAntiAffinity": strategies.RemovePodsViolatingInterPodAntiAffinity,
+		"RemovePodsViolatingNodeAffinity":         strategies.RemovePodsViolatingNodeAffinity,
+		"RemovePodsViolatingNodeTaints":           strategies.RemovePodsViolatingNodeTaints,
+		"RemovePodsHavingTooManyRestarts":         strategies.RemovePodsHavingTooManyRestarts,
+		"PodLifeTime":                             strategies.PodLifeTime,
 	}
 
-	if len(nodes) <= 1 {
-		klog.V(1).Infof("The cluster size is 0 or 1 meaning eviction causes service disruption or degradation. So aborting..")
-		return nil
-	}
-
-	nodePodCount := utils.InitializeNodePodCount(nodes)
 	wait.Until(func() {
-		strategies.RemoveDuplicatePods(rs, deschedulerPolicy.Strategies["RemoveDuplicates"], evictionPolicyGroupVersion, nodes, nodePodCount)
-		strategies.LowNodeUtilization(rs, deschedulerPolicy.Strategies["LowNodeUtilization"], evictionPolicyGroupVersion, nodes, nodePodCount)
-		strategies.RemovePodsViolatingInterPodAntiAffinity(rs, deschedulerPolicy.Strategies["RemovePodsViolatingInterPodAntiAffinity"], evictionPolicyGroupVersion, nodes, nodePodCount)
-		strategies.RemovePodsViolatingNodeAffinity(rs, deschedulerPolicy.Strategies["RemovePodsViolatingNodeAffinity"], evictionPolicyGroupVersion, nodes, nodePodCount)
-		strategies.RemovePodsViolatingNodeTaints(rs, deschedulerPolicy.Strategies["RemovePodsViolatingNodeTaints"], evictionPolicyGroupVersion, nodes, nodePodCount)
+		nodes, err := nodeutil.ReadyNodes(ctx, rs.Client, nodeInformer, rs.NodeSelector, stopChannel)
+		if err != nil {
+			klog.V(1).Infof("Unable to get ready nodes: %v", err)
+			close(stopChannel)
+			return
+		}
+
+		if len(nodes) <= 1 {
+			klog.V(1).Infof("The cluster size is 0 or 1 meaning eviction causes service disruption or degradation. So aborting..")
+			close(stopChannel)
+			return
+		}
+
+		podEvictor := evictions.NewPodEvictor(
+			rs.Client,
+			evictionPolicyGroupVersion,
+			rs.DryRun,
+			rs.MaxNoOfPodsToEvictPerNode,
+			nodes,
+		)
+
+		for name, f := range strategyFuncs {
+			if strategy := deschedulerPolicy.Strategies[api.StrategyName(name)]; strategy.Enabled {
+				f(ctx, rs.Client, strategy, nodes, rs.EvictLocalStoragePods, podEvictor)
+			}
+		}
 
 		// 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 {

pkg/descheduler/descheduler_test.go

@@ -0,0 +1,94 @@
+package descheduler
+
+import (
+	"context"
+	"fmt"
+	"strings"
+	"testing"
+	"time"
+
+	"k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/util/wait"
+	fakeclientset "k8s.io/client-go/kubernetes/fake"
+	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
+	"sigs.k8s.io/descheduler/pkg/api"
+	"sigs.k8s.io/descheduler/test"
+)
+
+func TestTaintsUpdated(t *testing.T) {
+	ctx := context.Background()
+	n1 := test.BuildTestNode("n1", 2000, 3000, 10, nil)
+	n2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
+
+	p1 := test.BuildTestPod(fmt.Sprintf("pod_1_%s", n1.Name), 200, 0, n1.Name, nil)
+	p1.ObjectMeta.OwnerReferences = []metav1.OwnerReference{
+		{},
+	}
+
+	client := fakeclientset.NewSimpleClientset(n1, n2, p1)
+	dp := &api.DeschedulerPolicy{
+		Strategies: api.StrategyList{
+			"RemovePodsViolatingNodeTaints": api.DeschedulerStrategy{
+				Enabled: true,
+			},
+		},
+	}
+
+	stopChannel := make(chan struct{})
+	defer close(stopChannel)
+
+	rs := options.NewDeschedulerServer()
+	rs.Client = client
+	rs.DeschedulingInterval = 100 * time.Millisecond
+	go func() {
+		err := RunDeschedulerStrategies(ctx, rs, dp, "v1beta1", stopChannel)
+		if err != nil {
+			t.Fatalf("Unable to run descheduler strategies: %v", err)
+		}
+	}()
+
+	// Wait for few cycles and then verify the only pod still exists
+	time.Sleep(300 * time.Millisecond)
+	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) < 1 {
+		t.Errorf("The pod was evicted before a node was tained")
+	}
+
+	n1WithTaint := n1.DeepCopy()
+	n1WithTaint.Spec.Taints = []v1.Taint{
+		{
+			Key:    "key",
+			Value:  "value",
+			Effect: v1.TaintEffectNoSchedule,
+		},
+	}
+
+	if _, err := client.CoreV1().Nodes().Update(ctx, n1WithTaint, metav1.UpdateOptions{}); err != nil {
+		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
+		}
+
+		return false, nil
+	}); err != nil {
+		t.Fatalf("Unable to evict pod, node taint did not get propagated to descheduler strategies")
+	}
+}

pkg/descheduler/evictions/evictions.go

@@ -17,6 +17,7 @@ limitations under the License.
 package evictions
 
 import (
+	"context"
 	"fmt"
 
 	v1 "k8s.io/api/core/v1"
@@ -32,9 +33,80 @@ import (
 	eutils "sigs.k8s.io/descheduler/pkg/descheduler/evictions/utils"
 )
 
-func EvictPod(client clientset.Interface, pod *v1.Pod, policyGroupVersion string, dryRun bool) (bool, error) {
+// nodePodEvictedCount keeps count of pods evicted on node
+type nodePodEvictedCount map[*v1.Node]int
+
+type PodEvictor struct {
+	client             clientset.Interface
+	policyGroupVersion string
+	dryRun             bool
+	maxPodsToEvict     int
+	nodepodCount       nodePodEvictedCount
+}
+
+func NewPodEvictor(
+	client clientset.Interface,
+	policyGroupVersion string,
+	dryRun bool,
+	maxPodsToEvict int,
+	nodes []*v1.Node,
+) *PodEvictor {
+	var nodePodCount = make(nodePodEvictedCount)
+	for _, node := range nodes {
+		// Initialize podsEvicted till now with 0.
+		nodePodCount[node] = 0
+	}
+
+	return &PodEvictor{
+		client:             client,
+		policyGroupVersion: policyGroupVersion,
+		dryRun:             dryRun,
+		maxPodsToEvict:     maxPodsToEvict,
+		nodepodCount:       nodePodCount,
+	}
+}
+
+// NodeEvicted gives a number of pods evicted for node
+func (pe *PodEvictor) NodeEvicted(node *v1.Node) int {
+	return pe.nodepodCount[node]
+}
+
+// TotalEvicted gives a number of pods evicted through all nodes
+func (pe *PodEvictor) TotalEvicted() int {
+	var total int
+	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 maxPodsToEvict 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) (bool, error) {
+	if pe.maxPodsToEvict > 0 && pe.nodepodCount[node]+1 > pe.maxPodsToEvict {
+		return false, fmt.Errorf("Maximum number %v of evicted pods per %q node reached", pe.maxPodsToEvict, node.Name)
+	}
+
+	err := evictPod(ctx, pe.client, pod, pe.policyGroupVersion, pe.dryRun)
+	if err != nil {
+		// err is used only for logging purposes
+		klog.Errorf("Error evicting pod: %#v in namespace %#v (%#v)", pod.Name, pod.Namespace, err)
+		return false, nil
+	}
+
+	pe.nodepodCount[node]++
+	if pe.dryRun {
+		klog.V(1).Infof("Evicted pod in dry run mode: %#v in namespace %#v", pod.Name, pod.Namespace)
+	} else {
+		klog.V(1).Infof("Evicted pod: %#v in namespace %#v", pod.Name, pod.Namespace)
+	}
+	return true, nil
+}
+
+func evictPod(ctx context.Context, client clientset.Interface, pod *v1.Pod, policyGroupVersion string, dryRun bool) error {
 	if dryRun {
-		return true, nil
+		return nil
 	}
 	deleteOptions := &metav1.DeleteOptions{}
 	// GracePeriodSeconds ?
@@ -49,7 +121,7 @@ func EvictPod(client clientset.Interface, pod *v1.Pod, policyGroupVersion string
 		},
 		DeleteOptions: deleteOptions,
 	}
-	err := client.PolicyV1beta1().Evictions(eviction.Namespace).Evict(eviction)
+	err := client.PolicyV1beta1().Evictions(eviction.Namespace).Evict(ctx, eviction)
 
 	if err == nil {
 		eventBroadcaster := record.NewBroadcaster()
@@ -57,12 +129,13 @@ func EvictPod(client clientset.Interface, pod *v1.Pod, policyGroupVersion string
 		eventBroadcaster.StartRecordingToSink(&clientcorev1.EventSinkImpl{Interface: client.CoreV1().Events(pod.Namespace)})
 		r := eventBroadcaster.NewRecorder(scheme.Scheme, v1.EventSource{Component: "sigs.k8s.io.descheduler"})
 		r.Event(pod, v1.EventTypeNormal, "Descheduled", "pod evicted by sigs.k8s.io/descheduler")
-		return true, nil
-	} else if apierrors.IsTooManyRequests(err) {
-		return false, fmt.Errorf("error when evicting pod (ignoring) %q: %v", pod.Name, err)
-	} else if apierrors.IsNotFound(err) {
-		return true, fmt.Errorf("pod not found when evicting %q: %v", pod.Name, err)
-	} else {
-		return false, err
+		return nil
 	}
+	if apierrors.IsTooManyRequests(err) {
+		return fmt.Errorf("error when evicting pod (ignoring) %q: %v", pod.Name, err)
+	}
+	if apierrors.IsNotFound(err) {
+		return fmt.Errorf("pod not found when evicting %q: %v", pod.Name, err)
+	}
+	return err
 }

pkg/descheduler/evictions/evictions_test.go

@@ -17,37 +17,40 @@ limitations under the License.
 package evictions
 
 import (
-	"k8s.io/api/core/v1"
+	"context"
+	"testing"
+
+	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/runtime"
 	"k8s.io/client-go/kubernetes/fake"
 	core "k8s.io/client-go/testing"
 	"sigs.k8s.io/descheduler/test"
-	"testing"
 )
 
 func TestEvictPod(t *testing.T) {
-	node1 := test.BuildTestNode("node1", 1000, 2000, 9)
-	pod1 := test.BuildTestPod("p1", 400, 0, "node1")
+	ctx := context.Background()
+	node1 := test.BuildTestNode("node1", 1000, 2000, 9, nil)
+	pod1 := test.BuildTestPod("p1", 400, 0, "node1", nil)
 	tests := []struct {
 		description string
 		node        *v1.Node
 		pod         *v1.Pod
 		pods        []v1.Pod
-		want        bool
+		want        error
 	}{
 		{
 			description: "test pod eviction - pod present",
 			node:        node1,
 			pod:         pod1,
 			pods:        []v1.Pod{*pod1},
-			want:        true,
+			want:        nil,
 		},
 		{
 			description: "test pod eviction - pod absent",
 			node:        node1,
 			pod:         pod1,
-			pods:        []v1.Pod{*test.BuildTestPod("p2", 400, 0, "node1"), *test.BuildTestPod("p3", 450, 0, "node1")},
-			want:        true,
+			pods:        []v1.Pod{*test.BuildTestPod("p2", 400, 0, "node1", nil), *test.BuildTestPod("p3", 450, 0, "node1", nil)},
+			want:        nil,
 		},
 	}
 
@@ -56,7 +59,7 @@ 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(fakeClient, test.pod, "v1", false)
+		got := evictPod(ctx, fakeClient, test.pod, "v1", false)
 		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)
 		}

pkg/descheduler/node/node.go

@@ -17,40 +17,34 @@ limitations under the License.
 package node
 
 import (
-	"time"
-
+	"context"
 	"k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/fields"
 	"k8s.io/apimachinery/pkg/labels"
+	coreinformers "k8s.io/client-go/informers/core/v1"
 	clientset "k8s.io/client-go/kubernetes"
-	corelisters "k8s.io/client-go/listers/core/v1"
-	"k8s.io/client-go/tools/cache"
 	"k8s.io/klog"
 	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
 // ReadyNodes returns ready nodes irrespective of whether they are
 // schedulable or not.
-func ReadyNodes(client clientset.Interface, nodeSelector string, stopChannel <-chan struct{}) ([]*v1.Node, error) {
+func ReadyNodes(ctx context.Context, client clientset.Interface, nodeInformer coreinformers.NodeInformer, nodeSelector string, stopChannel <-chan struct{}) ([]*v1.Node, error) {
 	ns, err := labels.Parse(nodeSelector)
 	if err != nil {
 		return []*v1.Node{}, err
 	}
 
 	var nodes []*v1.Node
-	nl := GetNodeLister(client, stopChannel)
-	if nl != nil {
-		// err is defined above
-		if nodes, err = nl.List(ns); err != nil {
-			return []*v1.Node{}, err
-		}
+	// err is defined above
+	if nodes, err = nodeInformer.Lister().List(ns); err != nil {
+		return []*v1.Node{}, err
 	}
 
 	if len(nodes) == 0 {
 		klog.V(2).Infof("node lister returned empty list, now fetch directly")
 
-		nItems, err := client.CoreV1().Nodes().List(metav1.ListOptions{LabelSelector: nodeSelector})
+		nItems, err := client.CoreV1().Nodes().List(ctx, metav1.ListOptions{LabelSelector: nodeSelector})
 		if err != nil {
 			return []*v1.Node{}, err
 		}
@@ -74,22 +68,6 @@ func ReadyNodes(client clientset.Interface, nodeSelector string, stopChannel <-c
 	return readyNodes, nil
 }
 
-func GetNodeLister(client clientset.Interface, stopChannel <-chan struct{}) corelisters.NodeLister {
-	if stopChannel == nil {
-		return nil
-	}
-	listWatcher := cache.NewListWatchFromClient(client.CoreV1().RESTClient(), "nodes", v1.NamespaceAll, fields.Everything())
-	store := cache.NewIndexer(cache.MetaNamespaceKeyFunc, cache.Indexers{cache.NamespaceIndex: cache.MetaNamespaceIndexFunc})
-	nodeLister := corelisters.NewNodeLister(store)
-	reflector := cache.NewReflector(listWatcher, &v1.Node{}, store, time.Hour)
-	go reflector.Run(stopChannel)
-
-	// To give some time so that listing works, chosen randomly
-	time.Sleep(100 * time.Millisecond)
-
-	return nodeLister
-}
-
 // IsReady checks if the descheduler could run against given node.
 func IsReady(node *v1.Node) bool {
 	for i := range node.Status.Conditions {
@@ -133,12 +111,9 @@ func PodFitsAnyNode(pod *v1.Pod, nodes []*v1.Node) bool {
 		if err != nil || !ok {
 			continue
 		}
-		if ok {
-			if !IsNodeUnschedulable(node) {
-				klog.V(2).Infof("Pod %v can possibly be scheduled on %v", pod.Name, node.Name)
-				return true
-			}
-			return false
+		if !IsNodeUnschedulable(node) {
+			klog.V(2).Infof("Pod %v can possibly be scheduled on %v", pod.Name, node.Name)
+			return true
 		}
 	}
 	return false

pkg/descheduler/node/node_test.go

@@ -17,23 +17,25 @@ limitations under the License.
 package node
 
 import (
+	"context"
 	"testing"
 
 	"k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/client-go/informers"
 	"k8s.io/client-go/kubernetes/fake"
 	"sigs.k8s.io/descheduler/test"
 )
 
 func TestReadyNodes(t *testing.T) {
-	node1 := test.BuildTestNode("node2", 1000, 2000, 9)
-	node2 := test.BuildTestNode("node3", 1000, 2000, 9)
+	node1 := test.BuildTestNode("node2", 1000, 2000, 9, nil)
+	node2 := test.BuildTestNode("node3", 1000, 2000, 9, nil)
 	node2.Status.Conditions = []v1.NodeCondition{{Type: v1.NodeMemoryPressure, Status: v1.ConditionTrue}}
-	node3 := test.BuildTestNode("node4", 1000, 2000, 9)
+	node3 := test.BuildTestNode("node4", 1000, 2000, 9, nil)
 	node3.Status.Conditions = []v1.NodeCondition{{Type: v1.NodeNetworkUnavailable, Status: v1.ConditionTrue}}
-	node4 := test.BuildTestNode("node5", 1000, 2000, 9)
+	node4 := test.BuildTestNode("node5", 1000, 2000, 9, nil)
 	node4.Spec.Unschedulable = true
-	node5 := test.BuildTestNode("node6", 1000, 2000, 9)
+	node5 := test.BuildTestNode("node6", 1000, 2000, 9, nil)
 	node5.Status.Conditions = []v1.NodeCondition{{Type: v1.NodeReady, Status: v1.ConditionFalse}}
 
 	if !IsReady(node1) {
@@ -55,14 +57,24 @@ func TestReadyNodes(t *testing.T) {
 }
 
 func TestReadyNodesWithNodeSelector(t *testing.T) {
-	node1 := test.BuildTestNode("node1", 1000, 2000, 9)
+	ctx := context.Background()
+	node1 := test.BuildTestNode("node1", 1000, 2000, 9, nil)
 	node1.Labels = map[string]string{"type": "compute"}
-	node2 := test.BuildTestNode("node2", 1000, 2000, 9)
+	node2 := test.BuildTestNode("node2", 1000, 2000, 9, nil)
 	node2.Labels = map[string]string{"type": "infra"}
 
 	fakeClient := fake.NewSimpleClientset(node1, node2)
 	nodeSelector := "type=compute"
-	nodes, _ := ReadyNodes(fakeClient, nodeSelector, nil)
+
+	sharedInformerFactory := informers.NewSharedInformerFactory(fakeClient, 0)
+	nodeInformer := sharedInformerFactory.Core().V1().Nodes()
+
+	stopChannel := make(chan struct{}, 0)
+	sharedInformerFactory.Start(stopChannel)
+	sharedInformerFactory.WaitForCacheSync(stopChannel)
+	defer close(stopChannel)
+
+	nodes, _ := ReadyNodes(ctx, fakeClient, nodeInformer, nodeSelector, nil)
 
 	if nodes[0].Name != "node1" {
 		t.Errorf("Expected node1, got %s", nodes[0].Name)
@@ -242,6 +254,49 @@ func TestPodFitsAnyNode(t *testing.T) {
 			},
 			success: true,
 		},
+		{
+			description: "Pod expected to fit one of the nodes (error node first)",
+			pod: &v1.Pod{
+				Spec: v1.PodSpec{
+					Affinity: &v1.Affinity{
+						NodeAffinity: &v1.NodeAffinity{
+							RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
+								NodeSelectorTerms: []v1.NodeSelectorTerm{
+									{
+										MatchExpressions: []v1.NodeSelectorRequirement{
+											{
+												Key:      nodeLabelKey,
+												Operator: "In",
+												Values: []string{
+													nodeLabelValue,
+												},
+											},
+										},
+									},
+								},
+							},
+						},
+					},
+				},
+			},
+			nodes: []*v1.Node{
+				{
+					ObjectMeta: metav1.ObjectMeta{
+						Labels: map[string]string{
+							nodeLabelKey: "no",
+						},
+					},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{
+						Labels: map[string]string{
+							nodeLabelKey: nodeLabelValue,
+						},
+					},
+				},
+			},
+			success: true,
+		},
 		{
 			description: "Pod expected to fit none of the nodes",
 			pod: &v1.Pod{

pkg/descheduler/pod/pods.go

@@ -17,6 +17,7 @@ limitations under the License.
 package pod
 
 import (
+	"context"
 	v1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/fields"
@@ -38,8 +39,8 @@ func IsEvictable(pod *v1.Pod, evictLocalStoragePods bool) bool {
 }
 
 // ListEvictablePodsOnNode returns the list of evictable pods on node.
-func ListEvictablePodsOnNode(client clientset.Interface, node *v1.Node, evictLocalStoragePods bool) ([]*v1.Pod, error) {
-	pods, err := ListPodsOnANode(client, node)
+func ListEvictablePodsOnNode(ctx context.Context, client clientset.Interface, node *v1.Node, evictLocalStoragePods bool) ([]*v1.Pod, error) {
+	pods, err := ListPodsOnANode(ctx, client, node)
 	if err != nil {
 		return []*v1.Pod{}, err
 	}
@@ -54,13 +55,13 @@ func ListEvictablePodsOnNode(client clientset.Interface, node *v1.Node, evictLoc
 	return evictablePods, nil
 }
 
-func ListPodsOnANode(client clientset.Interface, node *v1.Node) ([]*v1.Pod, error) {
+func ListPodsOnANode(ctx context.Context, client clientset.Interface, node *v1.Node) ([]*v1.Pod, error) {
 	fieldSelector, err := fields.ParseSelector("spec.nodeName=" + node.Name + ",status.phase!=" + string(v1.PodSucceeded) + ",status.phase!=" + string(v1.PodFailed))
 	if err != nil {
 		return []*v1.Pod{}, err
 	}
 
-	podList, err := client.CoreV1().Pods(v1.NamespaceAll).List(
+	podList, err := client.CoreV1().Pods(v1.NamespaceAll).List(ctx,
 		metav1.ListOptions{FieldSelector: fieldSelector.String()})
 	if err != nil {
 		return []*v1.Pod{}, err

pkg/descheduler/pod/pods_test.go

@@ -26,7 +26,7 @@ import (
 )
 
 func TestIsEvictable(t *testing.T) {
-	n1 := test.BuildTestNode("node1", 1000, 2000, 13)
+	n1 := test.BuildTestNode("node1", 1000, 2000, 13, nil)
 	type testCase struct {
 		pod                   *v1.Pod
 		runBefore             func(*v1.Pod)
@@ -36,14 +36,14 @@ func TestIsEvictable(t *testing.T) {
 
 	testCases := []testCase{
 		{
-			pod: test.BuildTestPod("p1", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p1", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 			},
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p2", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p2", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
 				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
@@ -51,14 +51,14 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p3", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p3", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
 			},
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p4", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p4", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
 				pod.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
@@ -66,7 +66,7 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p5", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p5", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 				pod.Spec.Volumes = []v1.Volume{
@@ -83,7 +83,7 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                false,
 		}, {
-			pod: test.BuildTestPod("p6", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p6", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 				pod.Spec.Volumes = []v1.Volume{
@@ -100,7 +100,7 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: true,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p7", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p7", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
 				pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
@@ -118,14 +118,14 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p8", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p8", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
 			},
 			evictLocalStoragePods: false,
 			result:                false,
 		}, {
-			pod: test.BuildTestPod("p9", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p9", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
 				pod.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
@@ -133,14 +133,14 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p10", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p10", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.Annotations = test.GetMirrorPodAnnotation()
 			},
 			evictLocalStoragePods: false,
 			result:                false,
 		}, {
-			pod: test.BuildTestPod("p11", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p11", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				pod.Annotations = test.GetMirrorPodAnnotation()
 				pod.Annotations["descheduler.alpha.kubernetes.io/evict"] = "true"
@@ -148,7 +148,7 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                true,
 		}, {
-			pod: test.BuildTestPod("p12", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p12", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				priority := utils.SystemCriticalPriority
 				pod.Spec.Priority = &priority
@@ -156,7 +156,7 @@ func TestIsEvictable(t *testing.T) {
 			evictLocalStoragePods: false,
 			result:                false,
 		}, {
-			pod: test.BuildTestPod("p13", 400, 0, n1.Name),
+			pod: test.BuildTestPod("p13", 400, 0, n1.Name, nil),
 			runBefore: func(pod *v1.Pod) {
 				priority := utils.SystemCriticalPriority
 				pod.Spec.Priority = &priority
@@ -179,15 +179,15 @@ func TestIsEvictable(t *testing.T) {
 	}
 }
 func TestPodTypes(t *testing.T) {
-	n1 := test.BuildTestNode("node1", 1000, 2000, 9)
-	p1 := test.BuildTestPod("p1", 400, 0, n1.Name)
+	n1 := test.BuildTestNode("node1", 1000, 2000, 9, nil)
+	p1 := test.BuildTestPod("p1", 400, 0, n1.Name, nil)
 
 	// These won't be evicted.
-	p2 := test.BuildTestPod("p2", 400, 0, n1.Name)
-	p3 := test.BuildTestPod("p3", 400, 0, n1.Name)
-	p4 := test.BuildTestPod("p4", 400, 0, n1.Name)
-	p5 := test.BuildTestPod("p5", 400, 0, n1.Name)
-	p6 := test.BuildTestPod("p6", 400, 0, n1.Name)
+	p2 := test.BuildTestPod("p2", 400, 0, n1.Name, nil)
+	p3 := test.BuildTestPod("p3", 400, 0, n1.Name, nil)
+	p4 := test.BuildTestPod("p4", 400, 0, n1.Name, nil)
+	p5 := test.BuildTestPod("p5", 400, 0, n1.Name, nil)
+	p6 := test.BuildTestPod("p6", 400, 0, n1.Name, nil)
 
 	p6.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 

pkg/descheduler/strategies/duplicates.go

@@ -17,81 +17,113 @@ limitations under the License.
 package strategies
 
 import (
+	"context"
+	"reflect"
+	"sort"
 	"strings"
 
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/util/sets"
 	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog"
 
-	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"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"
 )
 
-//type creator string
-type DuplicatePodsMap map[string][]*v1.Pod
-
 // RemoveDuplicatePods removes the duplicate pods on node. This strategy evicts all duplicate pods on node.
 // A pod is said to be a duplicate of other if both of them are from same creator, kind and are within the same
-// namespace. As of now, this strategy won't evict daemonsets, mirror pods, critical pods and pods with local storages.
-func RemoveDuplicatePods(ds *options.DeschedulerServer, strategy api.DeschedulerStrategy, policyGroupVersion string, nodes []*v1.Node, nodepodCount utils.NodePodEvictedCount) {
-	if !strategy.Enabled {
-		return
-	}
-	deleteDuplicatePods(ds.Client, policyGroupVersion, nodes, ds.DryRun, nodepodCount, ds.MaxNoOfPodsToEvictPerNode, ds.EvictLocalStoragePods)
-}
-
-// deleteDuplicatePods evicts the pod from node and returns the count of evicted pods.
-func deleteDuplicatePods(client clientset.Interface, policyGroupVersion string, nodes []*v1.Node, dryRun bool, nodepodCount utils.NodePodEvictedCount, maxPodsToEvict int, evictLocalStoragePods bool) int {
-	podsEvicted := 0
+// namespace, and have at least one container with the same image.
+// As of now, this strategy won't evict daemonsets, mirror pods, critical pods and pods with local storages.
+func RemoveDuplicatePods(
+	ctx context.Context,
+	client clientset.Interface,
+	strategy api.DeschedulerStrategy,
+	nodes []*v1.Node,
+	evictLocalStoragePods bool,
+	podEvictor *evictions.PodEvictor,
+) {
 	for _, node := range nodes {
 		klog.V(1).Infof("Processing node: %#v", node.Name)
-		dpm := ListDuplicatePodsOnANode(client, node, evictLocalStoragePods)
-		for creator, pods := range dpm {
-			if len(pods) > 1 {
-				klog.V(1).Infof("%#v", creator)
-				// i = 0 does not evict the first pod
-				for i := 1; i < len(pods); i++ {
-					if maxPodsToEvict > 0 && nodepodCount[node]+1 > maxPodsToEvict {
-						break
-					}
-					success, err := evictions.EvictPod(client, pods[i], policyGroupVersion, dryRun)
-					if !success {
-						klog.Infof("Error when evicting pod: %#v (%#v)", pods[i].Name, err)
-					} else {
-						nodepodCount[node]++
-						klog.V(1).Infof("Evicted pod: %#v (%#v)", pods[i].Name, err)
-					}
-				}
+		duplicatePods := listDuplicatePodsOnANode(ctx, client, node, strategy, evictLocalStoragePods)
+		for _, pod := range duplicatePods {
+			if _, err := podEvictor.EvictPod(ctx, pod, node); err != nil {
+				klog.Errorf("Error evicting pod: (%#v)", err)
+				break
 			}
 		}
-		podsEvicted += nodepodCount[node]
 	}
-	return podsEvicted
 }
 
-// ListDuplicatePodsOnANode lists duplicate pods on a given node.
-func ListDuplicatePodsOnANode(client clientset.Interface, node *v1.Node, evictLocalStoragePods bool) DuplicatePodsMap {
-	pods, err := podutil.ListEvictablePodsOnNode(client, node, evictLocalStoragePods)
+// listDuplicatePodsOnANode lists duplicate pods on a given node.
+// It checks for pods which have the same owner and have at least 1 container with the same image spec
+func listDuplicatePodsOnANode(ctx context.Context, client clientset.Interface, node *v1.Node, strategy api.DeschedulerStrategy, evictLocalStoragePods bool) []*v1.Pod {
+	pods, err := podutil.ListEvictablePodsOnNode(ctx, client, node, evictLocalStoragePods)
 	if err != nil {
 		return nil
 	}
-	return FindDuplicatePods(pods)
-}
 
-// FindDuplicatePods takes a list of pods and returns a duplicatePodsMap.
-func FindDuplicatePods(pods []*v1.Pod) DuplicatePodsMap {
-	dpm := DuplicatePodsMap{}
-	// Ignoring the error here as in the ListDuplicatePodsOnNode function we call ListEvictablePodsOnNode which checks for error.
+	duplicatePods := make([]*v1.Pod, 0, len(pods))
+	// Each pod has a list of owners and a list of containers, and each container has 1 image spec.
+	// For each pod, we go through all the OwnerRef/Image mappings and represent them as a "key" string.
+	// All of those mappings together makes a list of "key" strings that essentially represent that pod's uniqueness.
+	// This list of keys representing a single pod is then sorted alphabetically.
+	// If any other pod has a list that matches that pod's list, those pods are undeniably duplicates for the following reasons:
+	//   - The 2 pods have the exact same ownerrefs
+	//   - The 2 pods have the exact same container images
+	//
+	// duplicateKeysMap maps the first Namespace/Kind/Name/Image in a pod's list to a 2D-slice of all the other lists where that is the first key
+	// (Since we sort each pod's list, we only need to key the map on the first entry in each list. Any pod that doesn't have
+	// the same first entry is clearly not a duplicate. This makes lookup quick and minimizes storage needed).
+	// If any of the existing lists for that first key matches the current pod's list, the current pod is a duplicate.
+	// If not, then we add this pod's list to the list of lists for that key.
+	duplicateKeysMap := map[string][][]string{}
 	for _, pod := range pods {
 		ownerRefList := podutil.OwnerRef(pod)
+		if hasExcludedOwnerRefKind(ownerRefList, strategy) {
+			continue
+		}
+		podContainerKeys := make([]string, 0, len(ownerRefList)*len(pod.Spec.Containers))
 		for _, ownerRef := range ownerRefList {
-			// Namespace/Kind/Name should be unique for the cluster.
-			s := strings.Join([]string{pod.ObjectMeta.Namespace, ownerRef.Kind, ownerRef.Name}, "/")
-			dpm[s] = append(dpm[s], pod)
+			for _, container := range pod.Spec.Containers {
+				// Namespace/Kind/Name should be unique for the cluster.
+				// We also consider the image, as 2 pods could have the same owner but serve different purposes
+				// So any non-unique Namespace/Kind/Name/Image pattern is a duplicate pod.
+				s := strings.Join([]string{pod.ObjectMeta.Namespace, ownerRef.Kind, ownerRef.Name, container.Image}, "/")
+				podContainerKeys = append(podContainerKeys, s)
+			}
+		}
+		sort.Strings(podContainerKeys)
+
+		// If there have been any other pods with the same first "key", look through all the lists to see if any match
+		if existing, ok := duplicateKeysMap[podContainerKeys[0]]; ok {
+			for _, keys := range existing {
+				if reflect.DeepEqual(keys, podContainerKeys) {
+					duplicatePods = append(duplicatePods, pod)
+					break
+				}
+				// Found no matches, add this list of keys to the list of lists that have the same first key
+				duplicateKeysMap[podContainerKeys[0]] = append(duplicateKeysMap[podContainerKeys[0]], podContainerKeys)
+			}
+		} else {
+			// This is the first pod we've seen that has this first "key" entry
+			duplicateKeysMap[podContainerKeys[0]] = [][]string{podContainerKeys}
 		}
 	}
-	return dpm
+	return duplicatePods
+}
+
+func hasExcludedOwnerRefKind(ownerRefs []metav1.OwnerReference, strategy api.DeschedulerStrategy) bool {
+	if strategy.Params.RemoveDuplicates == nil {
+		return false
+	}
+	exclude := sets.NewString(strategy.Params.RemoveDuplicates.ExcludeOwnerKinds...)
+	for _, owner := range ownerRefs {
+		if exclude.Has(owner.Kind) {
+			return true
+		}
+	}
+	return false
 }

pkg/descheduler/strategies/duplicates_test.go

@@ -17,6 +17,7 @@ limitations under the License.
 package strategies
 
 import (
+	"context"
 	"testing"
 
 	"k8s.io/api/core/v1"
@@ -24,30 +25,41 @@ import (
 	"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 TestFindDuplicatePods(t *testing.T) {
+	ctx := context.Background()
 	// first setup pods
-	node := test.BuildTestNode("n1", 2000, 3000, 10)
-	p1 := test.BuildTestPod("p1", 100, 0, node.Name)
+	node := test.BuildTestNode("n1", 2000, 3000, 10, nil)
+	p1 := test.BuildTestPod("p1", 100, 0, node.Name, nil)
 	p1.Namespace = "dev"
-	p2 := test.BuildTestPod("p2", 100, 0, node.Name)
+	p2 := test.BuildTestPod("p2", 100, 0, node.Name, nil)
 	p2.Namespace = "dev"
-	p3 := test.BuildTestPod("p3", 100, 0, node.Name)
+	p3 := test.BuildTestPod("p3", 100, 0, node.Name, nil)
 	p3.Namespace = "dev"
-	p4 := test.BuildTestPod("p4", 100, 0, node.Name)
-	p5 := test.BuildTestPod("p5", 100, 0, node.Name)
-	p6 := test.BuildTestPod("p6", 100, 0, node.Name)
-	p7 := test.BuildTestPod("p7", 100, 0, node.Name)
+	p4 := test.BuildTestPod("p4", 100, 0, node.Name, nil)
+	p5 := test.BuildTestPod("p5", 100, 0, node.Name, nil)
+	p6 := test.BuildTestPod("p6", 100, 0, node.Name, nil)
+	p7 := test.BuildTestPod("p7", 100, 0, node.Name, nil)
 	p7.Namespace = "kube-system"
-	p8 := test.BuildTestPod("p8", 100, 0, node.Name)
+	p8 := test.BuildTestPod("p8", 100, 0, node.Name, nil)
 	p8.Namespace = "test"
-	p9 := test.BuildTestPod("p9", 100, 0, node.Name)
+	p9 := test.BuildTestPod("p9", 100, 0, node.Name, nil)
 	p9.Namespace = "test"
-	p10 := test.BuildTestPod("p10", 100, 0, node.Name)
+	p10 := test.BuildTestPod("p10", 100, 0, node.Name, nil)
 	p10.Namespace = "test"
+	p11 := test.BuildTestPod("p11", 100, 0, node.Name, nil)
+	p11.Namespace = "different-images"
+	p12 := test.BuildTestPod("p12", 100, 0, node.Name, nil)
+	p12.Namespace = "different-images"
+	p13 := test.BuildTestPod("p13", 100, 0, node.Name, nil)
+	p13.Namespace = "different-images"
+	p14 := test.BuildTestPod("p14", 100, 0, node.Name, nil)
+	p14.Namespace = "different-images"
 
 	// ### Evictable Pods ###
 
@@ -88,48 +100,92 @@ func TestFindDuplicatePods(t *testing.T) {
 	priority := utils.SystemCriticalPriority
 	p7.Spec.Priority = &priority
 
+	// Same owners, but different images
+	p11.Spec.Containers[0].Image = "foo"
+	p11.ObjectMeta.OwnerReferences = ownerRef1
+	p12.Spec.Containers[0].Image = "bar"
+	p12.ObjectMeta.OwnerReferences = ownerRef1
+
+	// Multiple containers
+	p13.ObjectMeta.OwnerReferences = ownerRef1
+	p13.Spec.Containers = append(p13.Spec.Containers, v1.Container{
+		Name:  "foo",
+		Image: "foo",
+	})
+
 	testCases := []struct {
 		description             string
 		maxPodsToEvict          int
 		pods                    []v1.Pod
 		expectedEvictedPodCount int
+		strategy                api.DeschedulerStrategy
 	}{
 		{
 			description:             "Three pods in the `dev` Namespace, bound to same ReplicaSet. 2 should be evicted.",
 			maxPodsToEvict:          5,
 			pods:                    []v1.Pod{*p1, *p2, *p3},
 			expectedEvictedPodCount: 2,
+			strategy:                api.DeschedulerStrategy{},
+		},
+		{
+			description:             "Three pods in the `dev` Namespace, bound to same ReplicaSet, but ReplicaSet kind is excluded. 0 should be evicted.",
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p1, *p2, *p3},
+			expectedEvictedPodCount: 0,
+			strategy:                api.DeschedulerStrategy{Params: api.StrategyParameters{RemoveDuplicates: &api.RemoveDuplicates{ExcludeOwnerKinds: []string{"ReplicaSet"}}}},
 		},
 		{
 			description:             "Three Pods in the `test` Namespace, bound to same ReplicaSet. 2 should be evicted.",
 			maxPodsToEvict:          5,
 			pods:                    []v1.Pod{*p8, *p9, *p10},
 			expectedEvictedPodCount: 2,
+			strategy:                api.DeschedulerStrategy{},
 		},
 		{
 			description:             "Three Pods in the `dev` Namespace, three Pods in the `test` Namespace. Bound to ReplicaSet with same name. 4 should be evicted.",
 			maxPodsToEvict:          5,
 			pods:                    []v1.Pod{*p1, *p2, *p3, *p8, *p9, *p10},
 			expectedEvictedPodCount: 4,
+			strategy:                api.DeschedulerStrategy{},
 		},
 		{
 			description:             "Pods are: part of DaemonSet, with local storage, mirror pod annotation, critical pod annotation - none should be evicted.",
 			maxPodsToEvict:          2,
 			pods:                    []v1.Pod{*p4, *p5, *p6, *p7},
 			expectedEvictedPodCount: 0,
+			strategy:                api.DeschedulerStrategy{},
 		},
 		{
 			description:             "Test all Pods: 4 should be evicted.",
 			maxPodsToEvict:          5,
 			pods:                    []v1.Pod{*p1, *p2, *p3, *p4, *p5, *p6, *p7, *p8, *p9, *p10},
 			expectedEvictedPodCount: 4,
+			strategy:                api.DeschedulerStrategy{},
+		},
+		{
+			description:             "Pods with the same owner but different images should not be evicted",
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p11, *p12},
+			expectedEvictedPodCount: 0,
+			strategy:                api.DeschedulerStrategy{},
+		},
+		{
+			description:             "Pods with multiple containers should not match themselves",
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p13},
+			expectedEvictedPodCount: 0,
+			strategy:                api.DeschedulerStrategy{},
+		},
+		{
+			description:             "Pods with matching ownerrefs and at not all matching image should not trigger an eviction",
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p11, *p13},
+			expectedEvictedPodCount: 0,
+			strategy:                api.DeschedulerStrategy{},
 		},
 	}
 
 	for _, testCase := range testCases {
-
-		npe := utils.NodePodEvictedCount{}
-		npe[node] = 0
 		fakeClient := &fake.Clientset{}
 		fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
 			return true, &v1.PodList{Items: testCase.pods}, nil
@@ -137,7 +193,16 @@ func TestFindDuplicatePods(t *testing.T) {
 		fakeClient.Fake.AddReactor("get", "nodes", func(action core.Action) (bool, runtime.Object, error) {
 			return true, node, nil
 		})
-		podsEvicted := deleteDuplicatePods(fakeClient, "v1", []*v1.Node{node}, false, npe, testCase.maxPodsToEvict, false)
+		podEvictor := evictions.NewPodEvictor(
+			fakeClient,
+			"v1",
+			false,
+			testCase.maxPodsToEvict,
+			[]*v1.Node{node},
+		)
+
+		RemoveDuplicatePods(ctx, fakeClient, testCase.strategy, []*v1.Node{node}, false, podEvictor)
+		podsEvicted := podEvictor.TotalEvicted()
 		if podsEvicted != testCase.expectedEvictedPodCount {
 			t.Errorf("Test error for description: %s. Expected evicted pods count %v, got %v", testCase.description, testCase.expectedEvictedPodCount, podsEvicted)
 		}

pkg/descheduler/strategies/lownodeutilization.go

@@ -17,13 +17,14 @@ limitations under the License.
 package strategies
 
 import (
+	"context"
 	"sort"
 
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/resource"
 	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog"
-	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
+
 	"sigs.k8s.io/descheduler/pkg/api"
 	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
 	nodeutil "sigs.k8s.io/descheduler/pkg/descheduler/node"
@@ -32,23 +33,23 @@ import (
 )
 
 type NodeUsageMap struct {
-	node             *v1.Node
-	usage            api.ResourceThresholds
-	allPods          []*v1.Pod
-	nonRemovablePods []*v1.Pod
-	bePods           []*v1.Pod
-	bPods            []*v1.Pod
-	gPods            []*v1.Pod
+	node    *v1.Node
+	usage   api.ResourceThresholds
+	allPods []*v1.Pod
 }
 
 type NodePodsMap map[*v1.Node][]*v1.Pod
 
-func LowNodeUtilization(ds *options.DeschedulerServer, strategy api.DeschedulerStrategy, evictionPolicyGroupVersion string, nodes []*v1.Node, nodepodCount utils.NodePodEvictedCount) {
+func LowNodeUtilization(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, evictLocalStoragePods bool, podEvictor *evictions.PodEvictor) {
 	if !strategy.Enabled {
 		return
 	}
 	// todo: move to config validation?
 	// 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 strategy.Params.NodeResourceUtilizationThresholds == nil {
+		klog.V(1).Infof("NodeResourceUtilizationThresholds not set")
+		return
+	}
 
 	thresholds := strategy.Params.NodeResourceUtilizationThresholds.Thresholds
 	if !validateThresholds(thresholds) {
@@ -59,8 +60,8 @@ func LowNodeUtilization(ds *options.DeschedulerServer, strategy api.DeschedulerS
 		return
 	}
 
-	npm := createNodePodsMap(ds.Client, nodes)
-	lowNodes, targetNodes := classifyNodes(npm, thresholds, targetThresholds, ds.EvictLocalStoragePods)
+	npm := createNodePodsMap(ctx, client, nodes)
+	lowNodes, targetNodes := classifyNodes(npm, thresholds, targetThresholds, evictLocalStoragePods)
 
 	klog.V(1).Infof("Criteria for a node under utilization: CPU: %v, Mem: %v, Pods: %v",
 		thresholds[v1.ResourceCPU], thresholds[v1.ResourceMemory], thresholds[v1.ResourcePods])
@@ -90,9 +91,15 @@ func LowNodeUtilization(ds *options.DeschedulerServer, strategy api.DeschedulerS
 		targetThresholds[v1.ResourceCPU], targetThresholds[v1.ResourceMemory], targetThresholds[v1.ResourcePods])
 	klog.V(1).Infof("Total number of nodes above target utilization: %v", len(targetNodes))
 
-	totalPodsEvicted := evictPodsFromTargetNodes(ds.Client, evictionPolicyGroupVersion, targetNodes, lowNodes, targetThresholds, ds.DryRun, ds.MaxNoOfPodsToEvictPerNode, nodepodCount)
-	klog.V(1).Infof("Total number of pods evicted: %v", totalPodsEvicted)
+	evictPodsFromTargetNodes(
+		ctx,
+		targetNodes,
+		lowNodes,
+		targetThresholds,
+		evictLocalStoragePods,
+		podEvictor)
 
+	klog.V(1).Infof("Total number of pods evicted: %v", podEvictor.TotalEvicted())
 }
 
 func validateThresholds(thresholds api.ResourceThresholds) bool {
@@ -133,9 +140,12 @@ func validateTargetThresholds(targetThresholds api.ResourceThresholds) bool {
 func classifyNodes(npm NodePodsMap, thresholds api.ResourceThresholds, targetThresholds api.ResourceThresholds, evictLocalStoragePods bool) ([]NodeUsageMap, []NodeUsageMap) {
 	lowNodes, targetNodes := []NodeUsageMap{}, []NodeUsageMap{}
 	for node, pods := range npm {
-		usage, allPods, nonRemovablePods, bePods, bPods, gPods := NodeUtilization(node, pods, evictLocalStoragePods)
-		nuMap := NodeUsageMap{node, usage, allPods, nonRemovablePods, bePods, bPods, gPods}
-
+		usage := nodeUtilization(node, pods, evictLocalStoragePods)
+		nuMap := NodeUsageMap{
+			node:    node,
+			usage:   usage,
+			allPods: pods,
+		}
 		// Check if node is underutilized and if we can schedule pods on it.
 		if !nodeutil.IsNodeUnschedulable(node) && IsNodeWithLowUtilization(usage, thresholds) {
 			klog.V(2).Infof("Node %#v is under utilized with usage: %#v", node.Name, usage)
@@ -146,7 +156,6 @@ func classifyNodes(npm NodePodsMap, thresholds api.ResourceThresholds, targetThr
 		} else {
 			klog.V(2).Infof("Node %#v is appropriately utilized with usage: %#v", node.Name, usage)
 		}
-		klog.V(2).Infof("allPods:%v, nonRemovablePods:%v, bePods:%v, bPods:%v, gPods:%v", len(allPods), len(nonRemovablePods), len(bePods), len(bPods), len(gPods))
 	}
 	return lowNodes, targetNodes
 }
@@ -154,14 +163,21 @@ func classifyNodes(npm NodePodsMap, thresholds api.ResourceThresholds, targetThr
 // evictPodsFromTargetNodes 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 evictPodsFromTargetNodes(client clientset.Interface, evictionPolicyGroupVersion string, targetNodes, lowNodes []NodeUsageMap, targetThresholds api.ResourceThresholds, dryRun bool, maxPodsToEvict int, nodepodCount utils.NodePodEvictedCount) int {
-	podsEvicted := 0
+func evictPodsFromTargetNodes(
+	ctx context.Context,
+	targetNodes, lowNodes []NodeUsageMap,
+	targetThresholds api.ResourceThresholds,
+	evictLocalStoragePods bool,
+	podEvictor *evictions.PodEvictor,
+) {
 
 	SortNodesByUsage(targetNodes)
 
 	// upper bound on total number of pods/cpu/memory to be moved
 	var totalPods, totalCPU, totalMem float64
+	var taintsOfLowNodes = make(map[string][]v1.Taint, len(lowNodes))
 	for _, node := range lowNodes {
+		taintsOfLowNodes[node.node.Name] = node.node.Spec.Taints
 		nodeCapacity := node.node.Status.Capacity
 		if len(node.node.Status.Allocatable) > 0 {
 			nodeCapacity = node.node.Status.Allocatable
@@ -192,62 +208,67 @@ func evictPodsFromTargetNodes(client clientset.Interface, evictionPolicyGroupVer
 			nodeCapacity = node.node.Status.Allocatable
 		}
 		klog.V(3).Infof("evicting pods from node %#v with usage: %#v", node.node.Name, node.usage)
-		currentPodsEvicted := nodepodCount[node.node]
+
+		nonRemovablePods, bestEffortPods, burstablePods, guaranteedPods := classifyPods(node.allPods, evictLocalStoragePods)
+		klog.V(2).Infof("allPods:%v, nonRemovablePods:%v, bestEffortPods:%v, burstablePods:%v, guaranteedPods:%v", len(node.allPods), len(nonRemovablePods), len(bestEffortPods), len(burstablePods), len(guaranteedPods))
 
 		// Check if one pod has priority, if yes, assume that all pods have priority and evict pods based on priority.
 		if node.allPods[0].Spec.Priority != nil {
 			klog.V(1).Infof("All pods have priority associated with them. Evicting pods based on priority")
 			evictablePods := make([]*v1.Pod, 0)
-			evictablePods = append(append(node.bPods, node.bePods...), node.gPods...)
+			evictablePods = append(append(burstablePods, bestEffortPods...), guaranteedPods...)
 
 			// 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.
 			sortPodsBasedOnPriority(evictablePods)
-			evictPods(evictablePods, client, evictionPolicyGroupVersion, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, &currentPodsEvicted, dryRun, maxPodsToEvict)
+			evictPods(ctx, evictablePods, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, taintsOfLowNodes, podEvictor, node.node)
 		} else {
 			// TODO: Remove this when we support only priority.
 			//  Falling back to evicting pods based on priority.
 			klog.V(1).Infof("Evicting pods based on QoS")
-			klog.V(1).Infof("There are %v non-evictable pods on the node", len(node.nonRemovablePods))
+			klog.V(1).Infof("There are %v non-evictable pods on the node", len(nonRemovablePods))
 			// evict best effort pods
-			evictPods(node.bePods, client, evictionPolicyGroupVersion, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, &currentPodsEvicted, dryRun, maxPodsToEvict)
+			evictPods(ctx, bestEffortPods, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, taintsOfLowNodes, podEvictor, node.node)
 			// evict burstable pods
-			evictPods(node.bPods, client, evictionPolicyGroupVersion, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, &currentPodsEvicted, dryRun, maxPodsToEvict)
+			evictPods(ctx, burstablePods, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, taintsOfLowNodes, podEvictor, node.node)
 			// evict guaranteed pods
-			evictPods(node.gPods, client, evictionPolicyGroupVersion, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, &currentPodsEvicted, dryRun, maxPodsToEvict)
+			evictPods(ctx, guaranteedPods, targetThresholds, nodeCapacity, node.usage, &totalPods, &totalCPU, &totalMem, taintsOfLowNodes, podEvictor, node.node)
 		}
-		nodepodCount[node.node] = currentPodsEvicted
-		podsEvicted = podsEvicted + nodepodCount[node.node]
-		klog.V(1).Infof("%v pods evicted from node %#v with usage %v", nodepodCount[node.node], node.node.Name, node.usage)
+		klog.V(1).Infof("%v pods evicted from node %#v with usage %v", podEvictor.NodeEvicted(node.node), node.node.Name, node.usage)
 	}
-	return podsEvicted
 }
 
-func evictPods(inputPods []*v1.Pod,
-	client clientset.Interface,
-	evictionPolicyGroupVersion string,
+func evictPods(
+	ctx context.Context,
+	inputPods []*v1.Pod,
 	targetThresholds api.ResourceThresholds,
 	nodeCapacity v1.ResourceList,
 	nodeUsage api.ResourceThresholds,
 	totalPods *float64,
 	totalCPU *float64,
 	totalMem *float64,
-	podsEvicted *int,
-	dryRun bool, maxPodsToEvict int) {
+	taintsOfLowNodes map[string][]v1.Taint,
+	podEvictor *evictions.PodEvictor,
+	node *v1.Node) {
 	if IsNodeAboveTargetUtilization(nodeUsage, targetThresholds) && (*totalPods > 0 || *totalCPU > 0 || *totalMem > 0) {
 		onePodPercentage := api.Percentage((float64(1) * 100) / float64(nodeCapacity.Pods().Value()))
 		for _, pod := range inputPods {
-			if maxPodsToEvict > 0 && *podsEvicted+1 > maxPodsToEvict {
-				break
+			if !utils.PodToleratesTaints(pod, taintsOfLowNodes) {
+				klog.V(3).Infof("Skipping eviction for Pod: %#v, doesn't tolerate node taint", pod.Name)
+				continue
 			}
+
 			cUsage := utils.GetResourceRequest(pod, v1.ResourceCPU)
 			mUsage := utils.GetResourceRequest(pod, v1.ResourceMemory)
-			success, err := evictions.EvictPod(client, pod, evictionPolicyGroupVersion, dryRun)
-			if !success {
-				klog.Warningf("Error when evicting pod: %#v (%#v)", pod.Name, err)
-			} else {
-				klog.V(3).Infof("Evicted pod: %#v (%#v)", pod.Name, err)
+
+			success, err := podEvictor.EvictPod(ctx, pod, node)
+			if err != nil {
+				klog.Errorf("Error evicting pod: (%#v)", err)
+				break
+			}
+
+			if success {
+				klog.V(3).Infof("Evicted pod: %#v", pod.Name)
 				// update remaining pods
-				*podsEvicted++
 				nodeUsage[v1.ResourcePods] -= onePodPercentage
 				*totalPods--
 
@@ -310,10 +331,10 @@ func sortPodsBasedOnPriority(evictablePods []*v1.Pod) {
 }
 
 // createNodePodsMap returns nodepodsmap with evictable pods on node.
-func createNodePodsMap(client clientset.Interface, nodes []*v1.Node) NodePodsMap {
+func createNodePodsMap(ctx context.Context, client clientset.Interface, nodes []*v1.Node) NodePodsMap {
 	npm := NodePodsMap{}
 	for _, node := range nodes {
-		pods, err := podutil.ListPodsOnANode(client, node)
+		pods, err := podutil.ListPodsOnANode(ctx, client, node)
 		if err != nil {
 			klog.Warningf("node %s will not be processed, error in accessing its pods (%#v)", node.Name, err)
 		} else {
@@ -349,38 +370,18 @@ func IsNodeWithLowUtilization(nodeThresholds api.ResourceThresholds, thresholds
 	return true
 }
 
-// NodeUtilization returns the current usage of node.
-func NodeUtilization(node *v1.Node, pods []*v1.Pod, evictLocalStoragePods bool) (api.ResourceThresholds, []*v1.Pod, []*v1.Pod, []*v1.Pod, []*v1.Pod, []*v1.Pod) {
-	bePods := []*v1.Pod{}
-	nonRemovablePods := []*v1.Pod{}
-	bPods := []*v1.Pod{}
-	gPods := []*v1.Pod{}
-	totalReqs := map[v1.ResourceName]resource.Quantity{}
+func nodeUtilization(node *v1.Node, pods []*v1.Pod, evictLocalStoragePods bool) api.ResourceThresholds {
+	totalReqs := map[v1.ResourceName]*resource.Quantity{
+		v1.ResourceCPU:    {},
+		v1.ResourceMemory: {},
+	}
 	for _, pod := range pods {
-		// We need to compute the usage of nonRemovablePods unless it is a best effort pod. So, cannot use podutil.ListEvictablePodsOnNode
-		if !podutil.IsEvictable(pod, evictLocalStoragePods) {
-			nonRemovablePods = append(nonRemovablePods, pod)
-			if podutil.IsBestEffortPod(pod) {
-				continue
-			}
-		} else if podutil.IsBestEffortPod(pod) {
-			bePods = append(bePods, pod)
-			continue
-		} else if podutil.IsBurstablePod(pod) {
-			bPods = append(bPods, pod)
-		} else {
-			gPods = append(gPods, pod)
-		}
-
 		req, _ := utils.PodRequestsAndLimits(pod)
 		for name, quantity := range req {
 			if name == v1.ResourceCPU || name == v1.ResourceMemory {
-				if value, ok := totalReqs[name]; !ok {
-					totalReqs[name] = quantity.DeepCopy()
-				} else {
-					value.Add(quantity)
-					totalReqs[name] = value
-				}
+				// 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)
 			}
 		}
 	}
@@ -390,12 +391,42 @@ func NodeUtilization(node *v1.Node, pods []*v1.Pod, evictLocalStoragePods bool)
 		nodeCapacity = node.Status.Allocatable
 	}
 
-	usage := api.ResourceThresholds{}
-	totalCPUReq := totalReqs[v1.ResourceCPU]
-	totalMemReq := totalReqs[v1.ResourceMemory]
 	totalPods := len(pods)
-	usage[v1.ResourceCPU] = api.Percentage((float64(totalCPUReq.MilliValue()) * 100) / float64(nodeCapacity.Cpu().MilliValue()))
-	usage[v1.ResourceMemory] = api.Percentage(float64(totalMemReq.Value()) / float64(nodeCapacity.Memory().Value()) * 100)
-	usage[v1.ResourcePods] = api.Percentage((float64(totalPods) * 100) / float64(nodeCapacity.Pods().Value()))
-	return usage, pods, nonRemovablePods, bePods, bPods, gPods
+	return api.ResourceThresholds{
+		v1.ResourceCPU:    api.Percentage((float64(totalReqs[v1.ResourceCPU].MilliValue()) * 100) / float64(nodeCapacity.Cpu().MilliValue())),
+		v1.ResourceMemory: api.Percentage(float64(totalReqs[v1.ResourceMemory].Value()) / float64(nodeCapacity.Memory().Value()) * 100),
+		v1.ResourcePods:   api.Percentage((float64(totalPods) * 100) / float64(nodeCapacity.Pods().Value())),
+	}
+}
+
+func classifyPods(pods []*v1.Pod, evictLocalStoragePods bool) ([]*v1.Pod, []*v1.Pod, []*v1.Pod, []*v1.Pod) {
+	var nonRemovablePods, bestEffortPods, burstablePods, guaranteedPods []*v1.Pod
+
+	// From https://kubernetes.io/docs/tasks/configure-pod-container/quality-service-pod/
+	//
+	// For a Pod to be given a QoS class of Guaranteed:
+	// - every Container in the Pod must have a memory limit and a memory request, and they must be the same.
+	// - every Container in the Pod must have a CPU limit and a CPU request, and they must be the same.
+	// A Pod is given a QoS class of Burstable if:
+	// - the Pod does not meet the criteria for QoS class Guaranteed.
+	// - at least one Container in the Pod has a memory or CPU request.
+	// For a Pod to be given a QoS class of BestEffort, the Containers in the Pod must not have any memory or CPU limits or requests.
+
+	for _, pod := range pods {
+		if !podutil.IsEvictable(pod, evictLocalStoragePods) {
+			nonRemovablePods = append(nonRemovablePods, pod)
+			continue
+		}
+
+		switch utils.GetPodQOS(pod) {
+		case v1.PodQOSGuaranteed:
+			guaranteedPods = append(guaranteedPods, pod)
+		case v1.PodQOSBurstable:
+			burstablePods = append(burstablePods, pod)
+		default: // alias v1.PodQOSBestEffort
+			bestEffortPods = append(bestEffortPods, pod)
+		}
+	}
+
+	return nonRemovablePods, bestEffortPods, burstablePods, guaranteedPods
 }

pkg/descheduler/strategies/lownodeutilization_test.go

@@ -17,255 +17,388 @@ limitations under the License.
 package strategies
 
 import (
+	"context"
 	"fmt"
 	"strings"
 	"testing"
 
 	"reflect"
 
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
+	"k8s.io/api/policy/v1beta1"
 	"k8s.io/apimachinery/pkg/api/resource"
+	"k8s.io/apimachinery/pkg/fields"
 	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/apimachinery/pkg/runtime/schema"
+	serializer "k8s.io/apimachinery/pkg/runtime/serializer"
+	"k8s.io/apimachinery/pkg/watch"
 	"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"
 )
 
-// TODO: Make this table driven.
-func TestLowNodeUtilizationWithoutPriority(t *testing.T) {
-	var thresholds = make(api.ResourceThresholds)
-	var targetThresholds = make(api.ResourceThresholds)
-	thresholds[v1.ResourceCPU] = 30
-	thresholds[v1.ResourcePods] = 30
-	targetThresholds[v1.ResourceCPU] = 50
-	targetThresholds[v1.ResourcePods] = 50
+var (
+	lowPriority  = int32(0)
+	highPriority = int32(10000)
+)
 
-	n1 := test.BuildTestNode("n1", 4000, 3000, 9)
-	n2 := test.BuildTestNode("n2", 4000, 3000, 10)
-	n3 := test.BuildTestNode("n3", 4000, 3000, 10)
-	// Making n3 node unschedulable so that it won't counted in lowUtilized nodes list.
-	n3.Spec.Unschedulable = true
-	p1 := test.BuildTestPod("p1", 400, 0, n1.Name)
-	p2 := test.BuildTestPod("p2", 400, 0, n1.Name)
-	p3 := test.BuildTestPod("p3", 400, 0, n1.Name)
-	p4 := test.BuildTestPod("p4", 400, 0, n1.Name)
-	p5 := test.BuildTestPod("p5", 400, 0, n1.Name)
-
-	// These won't be evicted.
-	p6 := test.BuildTestPod("p6", 400, 0, n1.Name)
-	p7 := test.BuildTestPod("p7", 400, 0, n1.Name)
-	p8 := test.BuildTestPod("p8", 400, 0, n1.Name)
-
-	p1.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p2.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p3.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p4.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p5.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	// The following 4 pods won't get evicted.
-	// A daemonset.
-	p6.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
-	// A pod with local storage.
-	p7.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p7.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.
-	p7.Annotations = test.GetMirrorPodAnnotation()
-	// A Critical Pod.
-	p8.Namespace = "kube-system"
-	priority := utils.SystemCriticalPriority
-	p8.Spec.Priority = &priority
-	p9 := test.BuildTestPod("p9", 400, 0, n1.Name)
-	p9.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	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: []v1.Pod{*p1, *p2, *p3, *p4, *p5, *p6, *p7, *p8}}, nil
-		}
-		if strings.Contains(fieldString, "n2") {
-			return true, &v1.PodList{Items: []v1.Pod{*p9}}, nil
-		}
-		if strings.Contains(fieldString, "n3") {
-			return true, &v1.PodList{Items: []v1.Pod{}}, 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)
-		switch getAction.GetName() {
-		case n1.Name:
-			return true, n1, nil
-		case n2.Name:
-			return true, n2, nil
-		case n3.Name:
-			return true, n3, nil
-		}
-		return true, nil, fmt.Errorf("Wrong node: %v", getAction.GetName())
-	})
-	expectedPodsEvicted := 3
-	npm := createNodePodsMap(fakeClient, []*v1.Node{n1, n2, n3})
-	lowNodes, targetNodes := classifyNodes(npm, thresholds, targetThresholds, false)
-	if len(lowNodes) != 1 {
-		t.Errorf("After ignoring unschedulable nodes, expected only one node to be under utilized.")
-	}
-	npe := utils.NodePodEvictedCount{}
-	npe[n1] = 0
-	npe[n2] = 0
-	npe[n3] = 0
-	podsEvicted := evictPodsFromTargetNodes(fakeClient, "v1", targetNodes, lowNodes, targetThresholds, false, 3, npe)
-	if expectedPodsEvicted != podsEvicted {
-		t.Errorf("Expected %#v pods to be evicted but %#v got evicted", expectedPodsEvicted, podsEvicted)
-	}
+func setRSOwnerRef(pod *v1.Pod)          { pod.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList() }
+func setDSOwnerRef(pod *v1.Pod)          { pod.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList() }
+func setNormalOwnerRef(pod *v1.Pod)      { pod.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList() }
+func setHighPriority(pod *v1.Pod)        { pod.Spec.Priority = &highPriority }
+func setLowPriority(pod *v1.Pod)         { pod.Spec.Priority = &lowPriority }
+func setNodeUnschedulable(node *v1.Node) { node.Spec.Unschedulable = true }
 
+func makeBestEffortPod(pod *v1.Pod) {
+	pod.Spec.Containers[0].Resources.Requests = nil
+	pod.Spec.Containers[0].Resources.Requests = nil
+	pod.Spec.Containers[0].Resources.Limits = nil
+	pod.Spec.Containers[0].Resources.Limits = nil
 }
 
-// TODO: Make this table driven.
-func TestLowNodeUtilizationWithPriorities(t *testing.T) {
-	var thresholds = make(api.ResourceThresholds)
-	var targetThresholds = make(api.ResourceThresholds)
-	thresholds[v1.ResourceCPU] = 30
-	thresholds[v1.ResourcePods] = 30
-	targetThresholds[v1.ResourceCPU] = 50
-	targetThresholds[v1.ResourcePods] = 50
-	lowPriority := int32(0)
-	highPriority := int32(10000)
-	n1 := test.BuildTestNode("n1", 4000, 3000, 9)
-	n2 := test.BuildTestNode("n2", 4000, 3000, 10)
-	n3 := test.BuildTestNode("n3", 4000, 3000, 10)
-	// Making n3 node unschedulable so that it won't counted in lowUtilized nodes list.
-	n3.Spec.Unschedulable = true
-	p1 := test.BuildTestPod("p1", 400, 0, n1.Name)
-	p1.Spec.Priority = &highPriority
-	p2 := test.BuildTestPod("p2", 400, 0, n1.Name)
-	p2.Spec.Priority = &highPriority
-	p3 := test.BuildTestPod("p3", 400, 0, n1.Name)
-	p3.Spec.Priority = &highPriority
-	p4 := test.BuildTestPod("p4", 400, 0, n1.Name)
-	p4.Spec.Priority = &highPriority
-	p5 := test.BuildTestPod("p5", 400, 0, n1.Name)
-	p5.Spec.Priority = &lowPriority
+func makeBurstablePod(pod *v1.Pod) {
+	pod.Spec.Containers[0].Resources.Limits = nil
+	pod.Spec.Containers[0].Resources.Limits = nil
+}
 
-	// These won't be evicted.
-	p6 := test.BuildTestPod("p6", 400, 0, n1.Name)
-	p6.Spec.Priority = &highPriority
-	p7 := test.BuildTestPod("p7", 400, 0, n1.Name)
-	p7.Spec.Priority = &lowPriority
-	p8 := test.BuildTestPod("p8", 400, 0, n1.Name)
-	p8.Spec.Priority = &lowPriority
+func makeGuaranteedPod(pod *v1.Pod) {
+	pod.Spec.Containers[0].Resources.Limits[v1.ResourceCPU] = pod.Spec.Containers[0].Resources.Requests[v1.ResourceCPU]
+	pod.Spec.Containers[0].Resources.Limits[v1.ResourceMemory] = pod.Spec.Containers[0].Resources.Requests[v1.ResourceMemory]
+}
 
-	p1.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p2.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p3.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p4.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	p5.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	// The following 4 pods won't get evicted.
-	// A daemonset.
-	p6.ObjectMeta.OwnerReferences = test.GetDaemonSetOwnerRefList()
-	// A pod with local storage.
-	p7.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p7.Spec.Volumes = []v1.Volume{
+func TestLowNodeUtilization(t *testing.T) {
+	ctx := context.Background()
+	n1NodeName := "n1"
+	n2NodeName := "n2"
+	n3NodeName := "n3"
+
+	testCases := []struct {
+		name                         string
+		thresholds, targetThresholds api.ResourceThresholds
+		nodes                        map[string]*v1.Node
+		pods                         map[string]*v1.PodList
+		expectedPodsEvicted          int
+		evictedPods                  []string
+	}{
 		{
-			Name: "sample",
-			VolumeSource: v1.VolumeSource{
-				HostPath: &v1.HostPathVolumeSource{Path: "somePath"},
-				EmptyDir: &v1.EmptyDirVolumeSource{
-					SizeLimit: resource.NewQuantity(int64(10), resource.BinarySI)},
+			name: "without priorities",
+			thresholds: api.ResourceThresholds{
+				v1.ResourceCPU:  30,
+				v1.ResourcePods: 30,
 			},
+			targetThresholds: api.ResourceThresholds{
+				v1.ResourceCPU:  50,
+				v1.ResourcePods: 50,
+			},
+			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, setNodeUnschedulable),
+			},
+			pods: map[string]*v1.PodList{
+				n1NodeName: {
+					Items: []v1.Pod{
+						*test.BuildTestPod("p1", 400, 0, n1NodeName, setRSOwnerRef),
+						*test.BuildTestPod("p2", 400, 0, n1NodeName, setRSOwnerRef),
+						*test.BuildTestPod("p3", 400, 0, n1NodeName, setRSOwnerRef),
+						*test.BuildTestPod("p4", 400, 0, n1NodeName, setRSOwnerRef),
+						*test.BuildTestPod("p5", 400, 0, n1NodeName, setRSOwnerRef),
+						// These won't be evicted.
+						*test.BuildTestPod("p6", 400, 0, n1NodeName, setDSOwnerRef),
+						*test.BuildTestPod("p7", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							// A pod with local storage.
+							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("p8", 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{
+						*test.BuildTestPod("p9", 400, 0, n1NodeName, setRSOwnerRef),
+					},
+				},
+				n3NodeName: {},
+			},
+			expectedPodsEvicted: 3,
+		},
+		{
+			name: "with priorities",
+			thresholds: api.ResourceThresholds{
+				v1.ResourceCPU:  30,
+				v1.ResourcePods: 30,
+			},
+			targetThresholds: api.ResourceThresholds{
+				v1.ResourceCPU:  50,
+				v1.ResourcePods: 50,
+			},
+			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, setNodeUnschedulable),
+			},
+			pods: map[string]*v1.PodList{
+				n1NodeName: {
+					Items: []v1.Pod{
+						*test.BuildTestPod("p1", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							setHighPriority(pod)
+						}),
+						*test.BuildTestPod("p2", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							setHighPriority(pod)
+						}),
+						*test.BuildTestPod("p3", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							setHighPriority(pod)
+						}),
+						*test.BuildTestPod("p4", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							setHighPriority(pod)
+						}),
+						*test.BuildTestPod("p5", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							setLowPriority(pod)
+						}),
+						// These won't be evicted.
+						*test.BuildTestPod("p6", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setDSOwnerRef(pod)
+							setHighPriority(pod)
+						}),
+						*test.BuildTestPod("p7", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							// A pod with local storage.
+							setNormalOwnerRef(pod)
+							setLowPriority(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("p8", 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{
+						*test.BuildTestPod("p9", 400, 0, n1NodeName, setRSOwnerRef),
+					},
+				},
+				n3NodeName: {},
+			},
+			expectedPodsEvicted: 3,
+		},
+		{
+			name: "without priorities evicting best-effort pods only",
+			thresholds: api.ResourceThresholds{
+				v1.ResourceCPU:  30,
+				v1.ResourcePods: 30,
+			},
+			targetThresholds: api.ResourceThresholds{
+				v1.ResourceCPU:  50,
+				v1.ResourcePods: 50,
+			},
+			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, 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) {
+							setRSOwnerRef(pod)
+							makeBestEffortPod(pod)
+						}),
+						*test.BuildTestPod("p2", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							makeBestEffortPod(pod)
+						}),
+						*test.BuildTestPod("p3", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+						}),
+						*test.BuildTestPod("p4", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							makeBestEffortPod(pod)
+						}),
+						*test.BuildTestPod("p5", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setRSOwnerRef(pod)
+							makeBestEffortPod(pod)
+						}),
+						// These won't be evicted.
+						*test.BuildTestPod("p6", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							setDSOwnerRef(pod)
+						}),
+						*test.BuildTestPod("p7", 400, 0, n1NodeName, func(pod *v1.Pod) {
+							// A pod with local storage.
+							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("p8", 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{
+						*test.BuildTestPod("p9", 400, 0, n1NodeName, setRSOwnerRef),
+					},
+				},
+				n3NodeName: {},
+			},
+			expectedPodsEvicted: 4,
+			evictedPods:         []string{"p1", "p2", "p4", "p5"},
 		},
 	}
-	// A Mirror Pod.
-	p7.Annotations = test.GetMirrorPodAnnotation()
-	// A Critical Pod.
-	p8.Namespace = "kube-system"
-	priority := utils.SystemCriticalPriority
-	p8.Spec.Priority = &priority
-	p9 := test.BuildTestPod("p9", 400, 0, n1.Name)
-	p9.ObjectMeta.OwnerReferences = test.GetReplicaSetOwnerRefList()
-	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: []v1.Pod{*p1, *p2, *p3, *p4, *p5, *p6, *p7, *p8}}, nil
-		}
-		if strings.Contains(fieldString, "n2") {
-			return true, &v1.PodList{Items: []v1.Pod{*p9}}, nil
-		}
-		if strings.Contains(fieldString, "n3") {
-			return true, &v1.PodList{Items: []v1.Pod{}}, 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)
-		switch getAction.GetName() {
-		case n1.Name:
-			return true, n1, nil
-		case n2.Name:
-			return true, n2, nil
-		case n3.Name:
-			return true, n3, nil
-		}
-		return true, nil, fmt.Errorf("Wrong node: %v", getAction.GetName())
-	})
-	expectedPodsEvicted := 3
-	npm := createNodePodsMap(fakeClient, []*v1.Node{n1, n2, n3})
-	lowNodes, targetNodes := classifyNodes(npm, thresholds, targetThresholds, false)
-	if len(lowNodes) != 1 {
-		t.Errorf("After ignoring unschedulable nodes, expected only one node to be under utilized.")
-	}
-	npe := utils.NodePodEvictedCount{}
-	npe[n1] = 0
-	npe[n2] = 0
-	npe[n3] = 0
-	podsEvicted := evictPodsFromTargetNodes(fakeClient, "v1", targetNodes, lowNodes, targetThresholds, false, 3, npe)
-	if expectedPodsEvicted != podsEvicted {
-		t.Errorf("Expected %#v pods to be evicted but %#v got evicted", expectedPodsEvicted, podsEvicted)
-	}
 
+	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)
+			}
+
+			npm := createNodePodsMap(ctx, fakeClient, nodes)
+			lowNodes, targetNodes := classifyNodes(npm, test.thresholds, test.targetThresholds, false)
+			if len(lowNodes) != 1 {
+				t.Errorf("After ignoring unschedulable nodes, expected only one node to be under utilized.")
+			}
+			podEvictor := evictions.NewPodEvictor(
+				fakeClient,
+				"v1",
+				false,
+				test.expectedPodsEvicted,
+				nodes,
+			)
+
+			evictPodsFromTargetNodes(ctx, targetNodes, lowNodes, test.targetThresholds, false, 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 TestSortPodsByPriority(t *testing.T) {
-	n1 := test.BuildTestNode("n1", 4000, 3000, 9)
-	lowPriority := int32(0)
-	highPriority := int32(10000)
-	p1 := test.BuildTestPod("p1", 400, 0, n1.Name)
-	p1.Spec.Priority = &lowPriority
+	n1 := test.BuildTestNode("n1", 4000, 3000, 9, nil)
+
+	p1 := test.BuildTestPod("p1", 400, 0, n1.Name, setLowPriority)
 
 	// BestEffort
-	p2 := test.BuildTestPod("p2", 400, 0, n1.Name)
-	p2.Spec.Priority = &highPriority
-
-	p2.Spec.Containers[0].Resources.Requests = nil
-	p2.Spec.Containers[0].Resources.Limits = nil
+	p2 := test.BuildTestPod("p2", 400, 0, n1.Name, func(pod *v1.Pod) {
+		setHighPriority(pod)
+		makeBestEffortPod(pod)
+	})
 
 	// Burstable
-	p3 := test.BuildTestPod("p3", 400, 0, n1.Name)
-	p3.Spec.Priority = &highPriority
+	p3 := test.BuildTestPod("p3", 400, 0, n1.Name, func(pod *v1.Pod) {
+		setHighPriority(pod)
+		makeBurstablePod(pod)
+	})
 
 	// Guaranteed
-	p4 := test.BuildTestPod("p4", 400, 100, n1.Name)
-	p4.Spec.Priority = &highPriority
-	p4.Spec.Containers[0].Resources.Limits[v1.ResourceCPU] = *resource.NewMilliQuantity(400, resource.DecimalSI)
-	p4.Spec.Containers[0].Resources.Limits[v1.ResourceMemory] = *resource.NewQuantity(100, resource.DecimalSI)
+	p4 := test.BuildTestPod("p4", 400, 100, n1.Name, func(pod *v1.Pod) {
+		setHighPriority(pod)
+		makeGuaranteedPod(pod)
+	})
 
 	// Best effort with nil priorities.
-	p5 := test.BuildTestPod("p5", 400, 100, n1.Name)
+	p5 := test.BuildTestPod("p5", 400, 100, n1.Name, makeBestEffortPod)
 	p5.Spec.Priority = nil
-	p6 := test.BuildTestPod("p6", 400, 100, n1.Name)
-	p6.Spec.Containers[0].Resources.Limits[v1.ResourceCPU] = *resource.NewMilliQuantity(400, resource.DecimalSI)
-	p6.Spec.Containers[0].Resources.Limits[v1.ResourceMemory] = *resource.NewQuantity(100, resource.DecimalSI)
+
+	p6 := test.BuildTestPod("p6", 400, 100, n1.Name, makeGuaranteedPod)
 	p6.Spec.Priority = nil
 
 	podList := []*v1.Pod{p4, p3, p2, p1, p6, p5}
@@ -327,3 +460,183 @@ func TestValidateThresholds(t *testing.T) {
 		}
 	}
 }
+
+func newFake(objects ...runtime.Object) *core.Fake {
+	scheme := runtime.NewScheme()
+	codecs := serializer.NewCodecFactory(scheme)
+	fake.AddToScheme(scheme)
+	o := core.NewObjectTracker(scheme, codecs.UniversalDecoder())
+	for _, obj := range objects {
+		if err := o.Add(obj); err != nil {
+			panic(err)
+		}
+	}
+
+	fakePtr := core.Fake{}
+	fakePtr.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
+		objs, err := o.List(
+			schema.GroupVersionResource{Group: "", Version: "v1", Resource: "pods"},
+			schema.GroupVersionKind{Group: "", Version: "v1", Kind: "Pod"},
+			action.GetNamespace(),
+		)
+		if err != nil {
+			return true, nil, err
+		}
+
+		obj := &v1.PodList{
+			Items: []v1.Pod{},
+		}
+		for _, pod := range objs.(*v1.PodList).Items {
+			podFieldSet := fields.Set(map[string]string{
+				"spec.nodeName": pod.Spec.NodeName,
+				"status.phase":  string(pod.Status.Phase),
+			})
+			match := action.(core.ListAction).GetListRestrictions().Fields.Matches(podFieldSet)
+			if !match {
+				continue
+			}
+			obj.Items = append(obj.Items, *pod.DeepCopy())
+		}
+		return true, obj, nil
+	})
+	fakePtr.AddReactor("*", "*", core.ObjectReaction(o))
+	fakePtr.AddWatchReactor("*", core.DefaultWatchReactor(watch.NewFake(), nil))
+
+	return &fakePtr
+}
+
+func TestWithTaints(t *testing.T) {
+	ctx := context.Background()
+	strategy := api.DeschedulerStrategy{
+		Enabled: true,
+		Params: api.StrategyParameters{
+			NodeResourceUtilizationThresholds: &api.NodeResourceUtilizationThresholds{
+				Thresholds: api.ResourceThresholds{
+					v1.ResourcePods: 20,
+				},
+				TargetThresholds: api.ResourceThresholds{
+					v1.ResourcePods: 70,
+				},
+			},
+		},
+	}
+
+	n1 := test.BuildTestNode("n1", 2000, 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, 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, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_2_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_3_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_4_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_5_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_6_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_7_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_8_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				// Node 2 pods
+				test.BuildTestPod(fmt.Sprintf("pod_9_%s", n2.Name), 200, 0, n2.Name, 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, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_2_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_3_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_4_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_5_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_6_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_7_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_8_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				// Node 3 pods
+				test.BuildTestPod(fmt.Sprintf("pod_9_%s", n3withTaints.Name), 200, 0, n3withTaints.Name, 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), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_2_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_3_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_4_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_5_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_6_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				test.BuildTestPod(fmt.Sprintf("pod_7_%s", n1.Name), 200, 0, n1.Name, setRSOwnerRef),
+				podThatToleratesTaint,
+				// Node 3 pods
+				test.BuildTestPod(fmt.Sprintf("pod_9_%s", n3withTaints.Name), 200, 0, n3withTaints.Name, 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)
+			}
+
+			fakePtr := newFake(objs...)
+			var evictionCounter int
+			fakePtr.PrependReactor("create", "pods", func(action core.Action) (bool, runtime.Object, error) {
+				if action.GetSubresource() != "eviction" || action.GetResource().Resource != "pods" {
+					return false, nil, nil
+				}
+				evictionCounter++
+				return true, nil, nil
+			})
+
+			podEvictor := evictions.NewPodEvictor(
+				&fake.Clientset{Fake: *fakePtr},
+				"policy/v1",
+				false,
+				item.evictionsExpected,
+				item.nodes,
+			)
+
+			LowNodeUtilization(ctx, &fake.Clientset{Fake: *fakePtr}, strategy, item.nodes, false, podEvictor)
+
+			if item.evictionsExpected != evictionCounter {
+				t.Errorf("Expected %v evictions, got %v", item.evictionsExpected, evictionCounter)
+			}
+		})
+	}
+}

pkg/descheduler/strategies/node_affinity.go

@@ -17,27 +17,19 @@ limitations under the License.
 package strategies
 
 import (
-	"k8s.io/api/core/v1"
-	"k8s.io/klog"
-	"sigs.k8s.io/descheduler/pkg/utils"
+	"context"
+
+	v1 "k8s.io/api/core/v1"
+	clientset "k8s.io/client-go/kubernetes"
+	"k8s.io/klog"
 
-	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
 	"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"
 )
 
-func RemovePodsViolatingNodeAffinity(ds *options.DeschedulerServer, strategy api.DeschedulerStrategy, evictionPolicyGroupVersion string, nodes []*v1.Node, nodePodCount utils.NodePodEvictedCount) {
-	removePodsViolatingNodeAffinityCount(ds, strategy, evictionPolicyGroupVersion, nodes, nodePodCount, ds.MaxNoOfPodsToEvictPerNode, ds.EvictLocalStoragePods)
-}
-
-func removePodsViolatingNodeAffinityCount(ds *options.DeschedulerServer, strategy api.DeschedulerStrategy, evictionPolicyGroupVersion string, nodes []*v1.Node, nodepodCount utils.NodePodEvictedCount, maxPodsToEvict int, evictLocalStoragePods bool) int {
-	evictedPodCount := 0
-	if !strategy.Enabled {
-		return evictedPodCount
-	}
-
+func RemovePodsViolatingNodeAffinity(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, evictLocalStoragePods bool, podEvictor *evictions.PodEvictor) {
 	for _, nodeAffinity := range strategy.Params.NodeAffinityType {
 		klog.V(2).Infof("Executing for nodeAffinityType: %v", nodeAffinity)
 
@@ -46,31 +38,26 @@ func removePodsViolatingNodeAffinityCount(ds *options.DeschedulerServer, strateg
 			for _, node := range nodes {
 				klog.V(1).Infof("Processing node: %#v\n", node.Name)
 
-				pods, err := podutil.ListEvictablePodsOnNode(ds.Client, node, evictLocalStoragePods)
+				pods, err := podutil.ListEvictablePodsOnNode(ctx, client, node, evictLocalStoragePods)
 				if err != nil {
 					klog.Errorf("failed to get pods from %v: %v", node.Name, err)
 				}
 
 				for _, pod := range pods {
-					if maxPodsToEvict > 0 && nodepodCount[node]+1 > maxPodsToEvict {
-						break
-					}
 					if pod.Spec.Affinity != nil && pod.Spec.Affinity.NodeAffinity != nil && pod.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution != nil {
-
 						if !nodeutil.PodFitsCurrentNode(pod, node) && nodeutil.PodFitsAnyNode(pod, nodes) {
 							klog.V(1).Infof("Evicting pod: %v", pod.Name)
-							evictions.EvictPod(ds.Client, pod, evictionPolicyGroupVersion, ds.DryRun)
-							nodepodCount[node]++
+							if _, err := podEvictor.EvictPod(ctx, pod, node); err != nil {
+								klog.Errorf("Error evicting pod: (%#v)", err)
+								break
+							}
 						}
 					}
 				}
-				evictedPodCount += nodepodCount[node]
 			}
 		default:
 			klog.Errorf("invalid nodeAffinityType: %v", nodeAffinity)
-			return evictedPodCount
 		}
 	}
-	klog.V(1).Infof("Evicted %v pods", evictedPodCount)
-	return evictedPodCount
+	klog.V(1).Infof("Evicted %v pods", podEvictor.TotalEvicted())
 }

pkg/descheduler/strategies/node_affinity_test.go

@@ -17,20 +17,20 @@ limitations under the License.
 package strategies
 
 import (
+	"context"
 	"testing"
 
 	"k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/runtime"
 	"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/utils"
+	"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{
@@ -42,17 +42,17 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 
 	nodeLabelKey := "kubernetes.io/desiredNode"
 	nodeLabelValue := "yes"
-	nodeWithLabels := test.BuildTestNode("nodeWithLabels", 2000, 3000, 10)
+	nodeWithLabels := test.BuildTestNode("nodeWithLabels", 2000, 3000, 10, nil)
 	nodeWithLabels.Labels[nodeLabelKey] = nodeLabelValue
 
-	nodeWithoutLabels := test.BuildTestNode("nodeWithoutLabels", 2000, 3000, 10)
+	nodeWithoutLabels := test.BuildTestNode("nodeWithoutLabels", 2000, 3000, 10, nil)
 
-	unschedulableNodeWithLabels := test.BuildTestNode("unschedulableNodeWithLabels", 2000, 3000, 10)
+	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)
+		podWithNodeAffinity := test.BuildTestPod("podWithNodeAffinity", 100, 0, node.Name, nil)
 		podWithNodeAffinity.Spec.Affinity = &v1.Affinity{
 			NodeAffinity: &v1.NodeAffinity{
 				RequiredDuringSchedulingIgnoredDuringExecution: &v1.NodeSelector{
@@ -73,8 +73,8 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			},
 		}
 
-		pod1 := test.BuildTestPod("pod1", 100, 0, node.Name)
-		pod2 := test.BuildTestPod("pod2", 100, 0, node.Name)
+		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()
@@ -93,25 +93,8 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 		pods                    []v1.Pod
 		strategy                api.DeschedulerStrategy
 		expectedEvictedPodCount int
-		npe                     utils.NodePodEvictedCount
 		maxPodsToEvict          int
 	}{
-		{
-			description: "Strategy disabled, should not evict any pods",
-			strategy: api.DeschedulerStrategy{
-				Enabled: false,
-				Params: api.StrategyParameters{
-					NodeAffinityType: []string{
-						"requiredDuringSchedulingIgnoredDuringExecution",
-					},
-				},
-			},
-			expectedEvictedPodCount: 0,
-			pods:                    addPodsToNode(nodeWithoutLabels),
-			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			npe:                     utils.NodePodEvictedCount{nodeWithoutLabels: 0, nodeWithLabels: 0},
-			maxPodsToEvict:          0,
-		},
 		{
 			description: "Invalid strategy type, should not evict any pods",
 			strategy: api.DeschedulerStrategy{
@@ -125,7 +108,6 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			expectedEvictedPodCount: 0,
 			pods:                    addPodsToNode(nodeWithoutLabels),
 			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			npe:                     utils.NodePodEvictedCount{nodeWithoutLabels: 0, nodeWithLabels: 0},
 			maxPodsToEvict:          0,
 		},
 		{
@@ -134,7 +116,6 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			expectedEvictedPodCount: 0,
 			pods:                    addPodsToNode(nodeWithLabels),
 			nodes:                   []*v1.Node{nodeWithLabels},
-			npe:                     utils.NodePodEvictedCount{nodeWithLabels: 0},
 			maxPodsToEvict:          0,
 		},
 		{
@@ -143,7 +124,6 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			strategy:                requiredDuringSchedulingIgnoredDuringExecutionStrategy,
 			pods:                    addPodsToNode(nodeWithoutLabels),
 			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			npe:                     utils.NodePodEvictedCount{nodeWithoutLabels: 0, nodeWithLabels: 0},
 			maxPodsToEvict:          0,
 		},
 		{
@@ -152,7 +132,6 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			strategy:                requiredDuringSchedulingIgnoredDuringExecutionStrategy,
 			pods:                    addPodsToNode(nodeWithoutLabels),
 			nodes:                   []*v1.Node{nodeWithoutLabels, nodeWithLabels},
-			npe:                     utils.NodePodEvictedCount{nodeWithoutLabels: 0, nodeWithLabels: 0},
 			maxPodsToEvict:          1,
 		},
 		{
@@ -161,7 +140,6 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			strategy:                requiredDuringSchedulingIgnoredDuringExecutionStrategy,
 			pods:                    addPodsToNode(nodeWithoutLabels),
 			nodes:                   []*v1.Node{nodeWithoutLabels, unschedulableNodeWithLabels},
-			npe:                     utils.NodePodEvictedCount{nodeWithoutLabels: 0, unschedulableNodeWithLabels: 0},
 			maxPodsToEvict:          0,
 		},
 	}
@@ -173,11 +151,16 @@ func TestRemovePodsViolatingNodeAffinity(t *testing.T) {
 			return true, &v1.PodList{Items: tc.pods}, nil
 		})
 
-		ds := options.DeschedulerServer{
-			Client: fakeClient,
-		}
+		podEvictor := evictions.NewPodEvictor(
+			fakeClient,
+			"v1",
+			false,
+			tc.maxPodsToEvict,
+			tc.nodes,
+		)
 
-		actualEvictedPodCount := removePodsViolatingNodeAffinityCount(&ds, tc.strategy, "v1", tc.nodes, tc.npe, tc.maxPodsToEvict, false)
+		RemovePodsViolatingNodeAffinity(ctx, fakeClient, tc.strategy, tc.nodes, false, 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

@@ -17,123 +17,40 @@ limitations under the License.
 package strategies
 
 import (
-	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
+	"context"
+
 	"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"
 
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
 	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog"
 )
 
-const (
-	TolerationOpExists v1.TolerationOperator = "Exists"
-	TolerationOpEqual  v1.TolerationOperator = "Equal"
-)
-
-// RemovePodsViolatingNodeTaints with elimination strategy
-func RemovePodsViolatingNodeTaints(ds *options.DeschedulerServer, strategy api.DeschedulerStrategy, policyGroupVersion string, nodes []*v1.Node, nodePodCount utils.NodePodEvictedCount) {
-	if !strategy.Enabled {
-		return
-	}
-	deletePodsViolatingNodeTaints(ds.Client, policyGroupVersion, nodes, ds.DryRun, nodePodCount, ds.MaxNoOfPodsToEvictPerNode, ds.EvictLocalStoragePods)
-}
-
-// deletePodsViolatingNodeTaints evicts pods on the node which violate NoSchedule Taints on nodes
-func deletePodsViolatingNodeTaints(client clientset.Interface, policyGroupVersion string, nodes []*v1.Node, dryRun bool, nodePodCount utils.NodePodEvictedCount, maxPodsToEvict int, evictLocalStoragePods bool) int {
-	podsEvicted := 0
+// 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, evictLocalStoragePods bool, podEvictor *evictions.PodEvictor) {
 	for _, node := range nodes {
 		klog.V(1).Infof("Processing node: %#v\n", node.Name)
-		pods, err := podutil.ListEvictablePodsOnNode(client, node, evictLocalStoragePods)
+		pods, err := podutil.ListEvictablePodsOnNode(ctx, client, node, evictLocalStoragePods)
 		if err != nil {
 			//no pods evicted as error encountered retrieving evictable Pods
-			return 0
+			return
 		}
 		totalPods := len(pods)
 		for i := 0; i < totalPods; i++ {
-			if maxPodsToEvict > 0 && nodePodCount[node]+1 > maxPodsToEvict {
-				break
-			}
-			if !checkPodsSatisfyTolerations(pods[i], node) {
-				success, err := evictions.EvictPod(client, pods[i], policyGroupVersion, dryRun)
-				if !success {
-					klog.Errorf("Error when evicting pod: %#v (%#v)\n", pods[i].Name, err)
-				} else {
-					nodePodCount[node]++
-					klog.V(1).Infof("Evicted pod: %#v (%#v)", pods[i].Name, err)
+			if !utils.TolerationsTolerateTaintsWithFilter(
+				pods[i].Spec.Tolerations,
+				node.Spec.Taints,
+				func(taint *v1.Taint) bool { return taint.Effect == v1.TaintEffectNoSchedule },
+			) {
+				klog.V(2).Infof("Not all taints with NoSchedule effect are tolerated after update for pod %v on node %v", pods[i].Name, node.Name)
+				if _, err := podEvictor.EvictPod(ctx, pods[i], node); err != nil {
+					klog.Errorf("Error evicting pod: (%#v)", err)
+					break
 				}
 			}
 		}
-		podsEvicted += nodePodCount[node]
-	}
-	return podsEvicted
-}
-
-// checkPodsSatisfyTolerations checks if the node's taints (NoSchedule) are still satisfied by pods' tolerations.
-func checkPodsSatisfyTolerations(pod *v1.Pod, node *v1.Node) bool {
-	tolerations := pod.Spec.Tolerations
-	taints := node.Spec.Taints
-	if len(taints) == 0 {
-		return true
-	}
-	noScheduleTaints := getNoScheduleTaints(taints)
-	if !allTaintsTolerated(noScheduleTaints, tolerations) {
-		klog.V(2).Infof("Not all taints are tolerated after update for Pod %v on node %v", pod.Name, node.Name)
-		return false
-	}
-	return true
-}
-
-// getNoScheduleTaints return a slice of NoSchedule taints from the a slice of taints that it receives.
-func getNoScheduleTaints(taints []v1.Taint) []v1.Taint {
-	result := []v1.Taint{}
-	for i := range taints {
-		if taints[i].Effect == v1.TaintEffectNoSchedule {
-			result = append(result, taints[i])
-		}
-	}
-	return result
-}
-
-//toleratesTaint returns true if a toleration tolerates a taint, or false otherwise
-func toleratesTaint(toleration *v1.Toleration, taint *v1.Taint) bool {
-
-	if (len(toleration.Key) > 0 && toleration.Key != taint.Key) ||
-		(len(toleration.Effect) > 0 && toleration.Effect != taint.Effect) {
-		return false
-	}
-	switch toleration.Operator {
-	// empty operator means Equal
-	case "", TolerationOpEqual:
-		return toleration.Value == taint.Value
-	case TolerationOpExists:
-		return true
-	default:
-		return false
 	}
 }
-
-// allTaintsTolerated returns true if all are tolerated, or false otherwise.
-func allTaintsTolerated(taints []v1.Taint, tolerations []v1.Toleration) bool {
-	if len(taints) == 0 {
-		return true
-	}
-	if len(tolerations) == 0 && len(taints) > 0 {
-		return false
-	}
-	for i := range taints {
-		tolerated := false
-		for j := range tolerations {
-			if toleratesTaint(&tolerations[j], &taints[i]) {
-				tolerated = true
-				break
-			}
-		}
-		if !tolerated {
-			return false
-		}
-	}
-	return true
-}

pkg/descheduler/strategies/node_taint_test.go

@@ -1,6 +1,7 @@
 package strategies
 
 import (
+	"context"
 	"fmt"
 	"testing"
 
@@ -9,6 +10,8 @@ import (
 	"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"
 )
@@ -41,18 +44,18 @@ func addTolerationToPod(pod *v1.Pod, key, value string, index int) *v1.Pod {
 }
 
 func TestDeletePodsViolatingNodeTaints(t *testing.T) {
-
-	node1 := test.BuildTestNode("n1", 2000, 3000, 10)
+	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)
+	node2 := test.BuildTestNode("n2", 2000, 3000, 10, nil)
 	node1 = addTaintsToNode(node2, "testingTaint", "testing", []int{1})
 
-	p1 := test.BuildTestPod("p1", 100, 0, node1.Name)
-	p2 := test.BuildTestPod("p2", 100, 0, node1.Name)
-	p3 := test.BuildTestPod("p3", 100, 0, node1.Name)
-	p4 := test.BuildTestPod("p4", 100, 0, node1.Name)
-	p5 := test.BuildTestPod("p5", 100, 0, node1.Name)
-	p6 := test.BuildTestPod("p6", 100, 0, node1.Name)
+	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()
@@ -60,11 +63,11 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 	p4.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 	p5.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 	p6.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
-	p7 := test.BuildTestPod("p7", 100, 0, node2.Name)
-	p8 := test.BuildTestPod("p8", 100, 0, node2.Name)
-	p9 := test.BuildTestPod("p9", 100, 0, node2.Name)
-	p10 := test.BuildTestPod("p10", 100, 0, node2.Name)
-	p11 := test.BuildTestPod("p11", 100, 0, node2.Name)
+	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()
 
 	// The following 4 pods won't get evicted.
@@ -99,7 +102,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 		nodes                   []*v1.Node
 		pods                    []v1.Pod
 		evictLocalStoragePods   bool
-		npe                     utils.NodePodEvictedCount
 		maxPodsToEvict          int
 		expectedEvictedPodCount int
 	}{
@@ -109,7 +111,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			pods:                    []v1.Pod{*p1, *p2, *p3},
 			nodes:                   []*v1.Node{node1},
 			evictLocalStoragePods:   false,
-			npe:                     utils.NodePodEvictedCount{node1: 0},
 			maxPodsToEvict:          0,
 			expectedEvictedPodCount: 1, //p2 gets evicted
 		},
@@ -118,7 +119,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			pods:                    []v1.Pod{*p1, *p3, *p4},
 			nodes:                   []*v1.Node{node1},
 			evictLocalStoragePods:   false,
-			npe:                     utils.NodePodEvictedCount{node1: 0},
 			maxPodsToEvict:          0,
 			expectedEvictedPodCount: 1, //p4 gets evicted
 		},
@@ -127,7 +127,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			pods:                    []v1.Pod{*p1, *p5, *p6},
 			nodes:                   []*v1.Node{node1},
 			evictLocalStoragePods:   false,
-			npe:                     utils.NodePodEvictedCount{node1: 0},
 			maxPodsToEvict:          1,
 			expectedEvictedPodCount: 1, //p5 or p6 gets evicted
 		},
@@ -136,7 +135,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			pods:                    []v1.Pod{*p7, *p8, *p9, *p10},
 			nodes:                   []*v1.Node{node2},
 			evictLocalStoragePods:   false,
-			npe:                     utils.NodePodEvictedCount{node2: 0},
 			maxPodsToEvict:          0,
 			expectedEvictedPodCount: 0,
 		},
@@ -145,7 +143,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			pods:                    []v1.Pod{*p7, *p8, *p9, *p10},
 			nodes:                   []*v1.Node{node2},
 			evictLocalStoragePods:   true,
-			npe:                     utils.NodePodEvictedCount{node2: 0},
 			maxPodsToEvict:          0,
 			expectedEvictedPodCount: 1,
 		},
@@ -154,7 +151,6 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			pods:                    []v1.Pod{*p7, *p8, *p10, *p11},
 			nodes:                   []*v1.Node{node2},
 			evictLocalStoragePods:   false,
-			npe:                     utils.NodePodEvictedCount{node2: 0},
 			maxPodsToEvict:          0,
 			expectedEvictedPodCount: 1,
 		},
@@ -168,7 +164,16 @@ func TestDeletePodsViolatingNodeTaints(t *testing.T) {
 			return true, &v1.PodList{Items: tc.pods}, nil
 		})
 
-		actualEvictedPodCount := deletePodsViolatingNodeTaints(fakeClient, "v1", tc.nodes, false, tc.npe, tc.maxPodsToEvict, tc.evictLocalStoragePods)
+		podEvictor := evictions.NewPodEvictor(
+			fakeClient,
+			"v1",
+			false,
+			tc.maxPodsToEvict,
+			tc.nodes,
+		)
+
+		RemovePodsViolatingNodeTaints(ctx, fakeClient, api.DeschedulerStrategy{}, tc.nodes, tc.evictLocalStoragePods, 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)
 		}
@@ -188,7 +193,7 @@ func TestToleratesTaint(t *testing.T) {
 			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: TolerationOpExists,
+				Operator: v1.TolerationOpExists,
 				Effect:   v1.TaintEffectNoSchedule,
 			},
 			taint: v1.Taint{
@@ -201,7 +206,7 @@ func TestToleratesTaint(t *testing.T) {
 			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: TolerationOpExists,
+				Operator: v1.TolerationOpExists,
 				Effect:   v1.TaintEffectNoSchedule,
 			},
 			taint: v1.Taint{
@@ -215,7 +220,7 @@ func TestToleratesTaint(t *testing.T) {
 			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: TolerationOpExists,
+				Operator: v1.TolerationOpExists,
 				Effect:   v1.TaintEffectNoSchedule,
 			},
 			taint: v1.Taint{
@@ -229,7 +234,7 @@ func TestToleratesTaint(t *testing.T) {
 			description: "toleration and taint have the same key, effect and value, and operator is Equal, expect tolerated",
 			toleration: v1.Toleration{
 				Key:      "foo",
-				Operator: TolerationOpEqual,
+				Operator: v1.TolerationOpEqual,
 				Value:    "bar",
 				Effect:   v1.TaintEffectNoSchedule,
 			},
@@ -244,7 +249,7 @@ func TestToleratesTaint(t *testing.T) {
 			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: TolerationOpEqual,
+				Operator: v1.TolerationOpEqual,
 				Value:    "value1",
 				Effect:   v1.TaintEffectNoSchedule,
 			},
@@ -259,7 +264,7 @@ func TestToleratesTaint(t *testing.T) {
 			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: TolerationOpEqual,
+				Operator: v1.TolerationOpEqual,
 				Value:    "bar",
 				Effect:   v1.TaintEffectNoSchedule,
 			},
@@ -272,27 +277,8 @@ func TestToleratesTaint(t *testing.T) {
 		},
 	}
 	for _, tc := range testCases {
-		if tolerated := toleratesTaint(&tc.toleration, &tc.taint); tc.expectTolerated != tolerated {
+		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())
 		}
 	}
 }
-
-func TestFilterNoExecuteTaints(t *testing.T) {
-	taints := []v1.Taint{
-		{
-			Key:    "one",
-			Value:  "one",
-			Effect: v1.TaintEffectNoExecute,
-		},
-		{
-			Key:    "two",
-			Value:  "two",
-			Effect: v1.TaintEffectNoSchedule,
-		},
-	}
-	taints = getNoScheduleTaints(taints)
-	if len(taints) != 1 || taints[0].Key != "two" {
-		t.Errorf("Filtering doesn't work. Got %v", taints)
-	}
-}

pkg/descheduler/strategies/pod_antiaffinity.go

@@ -17,7 +17,7 @@ limitations under the License.
 package strategies
 
 import (
-	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
+	"context"
 	"sigs.k8s.io/descheduler/pkg/api"
 	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
 	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
@@ -29,35 +29,25 @@ import (
 	"k8s.io/klog"
 )
 
-// RemovePodsViolatingInterPodAntiAffinity with elimination strategy
-func RemovePodsViolatingInterPodAntiAffinity(ds *options.DeschedulerServer, strategy api.DeschedulerStrategy, policyGroupVersion string, nodes []*v1.Node, nodePodCount utils.NodePodEvictedCount) {
-	if !strategy.Enabled {
-		return
-	}
-	removePodsWithAffinityRules(ds.Client, policyGroupVersion, nodes, ds.DryRun, nodePodCount, ds.MaxNoOfPodsToEvictPerNode, ds.EvictLocalStoragePods)
-}
-
-// removePodsWithAffinityRules evicts pods on the node which are having a pod affinity rules.
-func removePodsWithAffinityRules(client clientset.Interface, policyGroupVersion string, nodes []*v1.Node, dryRun bool, nodePodCount utils.NodePodEvictedCount, maxPodsToEvict int, evictLocalStoragePods bool) int {
-	podsEvicted := 0
+// RemovePodsViolatingInterPodAntiAffinity evicts pods on the node which are having a pod affinity rules.
+func RemovePodsViolatingInterPodAntiAffinity(ctx context.Context, client clientset.Interface, strategy api.DeschedulerStrategy, nodes []*v1.Node, evictLocalStoragePods bool, podEvictor *evictions.PodEvictor) {
 	for _, node := range nodes {
 		klog.V(1).Infof("Processing node: %#v\n", node.Name)
-		pods, err := podutil.ListEvictablePodsOnNode(client, node, evictLocalStoragePods)
+		pods, err := podutil.ListEvictablePodsOnNode(ctx, client, node, evictLocalStoragePods)
 		if err != nil {
-			return 0
+			return
 		}
 		totalPods := len(pods)
 		for i := 0; i < totalPods; i++ {
-			if maxPodsToEvict > 0 && nodePodCount[node]+1 > maxPodsToEvict {
-				break
-			}
 			if checkPodsWithAntiAffinityExist(pods[i], pods) {
-				success, err := evictions.EvictPod(client, pods[i], policyGroupVersion, dryRun)
-				if !success {
-					klog.Infof("Error when evicting pod: %#v (%#v)\n", pods[i].Name, err)
-				} else {
-					nodePodCount[node]++
-					klog.V(1).Infof("Evicted pod: %#v (%#v)\n because of existing anti-affinity", pods[i].Name, err)
+				success, err := podEvictor.EvictPod(ctx, pods[i], node)
+				if err != nil {
+					klog.Errorf("Error evicting pod: (%#v)", err)
+					break
+				}
+
+				if success {
+					klog.V(1).Infof("Evicted pod: %#v\n because of existing anti-affinity", pods[i].Name)
 					// Since the current pod is evicted all other pods which have anti-affinity with this
 					// pod need not be evicted.
 					// Update pods.
@@ -67,9 +57,7 @@ func removePodsWithAffinityRules(client clientset.Interface, policyGroupVersion
 				}
 			}
 		}
-		podsEvicted += nodePodCount[node]
 	}
-	return podsEvicted
 }
 
 // checkPodsWithAntiAffinityExist checks if there are other pods on the node that the current pod cannot tolerate.

pkg/descheduler/strategies/pod_antiaffinity_test.go

@@ -17,6 +17,7 @@ limitations under the License.
 package strategies
 
 import (
+	"context"
 	"testing"
 
 	"k8s.io/api/core/v1"
@@ -24,16 +25,18 @@ import (
 	"k8s.io/apimachinery/pkg/runtime"
 	"k8s.io/client-go/kubernetes/fake"
 	core "k8s.io/client-go/testing"
-	"sigs.k8s.io/descheduler/pkg/utils"
+	"sigs.k8s.io/descheduler/pkg/api"
+	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
 	"sigs.k8s.io/descheduler/test"
 )
 
 func TestPodAntiAffinity(t *testing.T) {
-	node := test.BuildTestNode("n1", 2000, 3000, 10)
-	p1 := test.BuildTestPod("p1", 100, 0, node.Name)
-	p2 := test.BuildTestPod("p2", 100, 0, node.Name)
-	p3 := test.BuildTestPod("p3", 100, 0, node.Name)
-	p4 := test.BuildTestPod("p4", 100, 0, node.Name)
+	ctx := context.Background()
+	node := test.BuildTestNode("n1", 2000, 3000, 10, nil)
+	p1 := test.BuildTestPod("p1", 100, 0, node.Name, nil)
+	p2 := test.BuildTestPod("p2", 100, 0, node.Name, nil)
+	p3 := test.BuildTestPod("p3", 100, 0, node.Name, nil)
+	p4 := test.BuildTestPod("p4", 100, 0, node.Name, nil)
 	p2.Labels = map[string]string{"foo": "bar"}
 	p1.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
 	p2.ObjectMeta.OwnerReferences = test.GetNormalPodOwnerRefList()
@@ -45,26 +48,49 @@ func TestPodAntiAffinity(t *testing.T) {
 	setPodAntiAffinity(p3)
 	setPodAntiAffinity(p4)
 
-	// create fake client
-	fakeClient := &fake.Clientset{}
-	fakeClient.Fake.AddReactor("list", "pods", func(action core.Action) (bool, runtime.Object, error) {
-		return true, &v1.PodList{Items: []v1.Pod{*p1, *p2, *p3, *p4}}, nil
-	})
-	fakeClient.Fake.AddReactor("get", "nodes", func(action core.Action) (bool, runtime.Object, error) {
-		return true, node, nil
-	})
-	npe := utils.NodePodEvictedCount{}
-	npe[node] = 0
-	expectedEvictedPodCount := 3
-	podsEvicted := removePodsWithAffinityRules(fakeClient, "v1", []*v1.Node{node}, false, npe, 0, false)
-	if podsEvicted != expectedEvictedPodCount {
-		t.Errorf("Unexpected no of pods evicted: pods evicted: %d, expected: %d", podsEvicted, expectedEvictedPodCount)
+	tests := []struct {
+		description             string
+		maxPodsToEvict          int
+		pods                    []v1.Pod
+		expectedEvictedPodCount int
+	}{
+		{
+			description:             "Maximum pods to evict - 0",
+			maxPodsToEvict:          0,
+			pods:                    []v1.Pod{*p1, *p2, *p3, *p4},
+			expectedEvictedPodCount: 3,
+		},
+		{
+			description:             "Maximum pods to evict - 3",
+			maxPodsToEvict:          3,
+			pods:                    []v1.Pod{*p1, *p2, *p3, *p4},
+			expectedEvictedPodCount: 3,
+		},
 	}
-	npe[node] = 0
-	expectedEvictedPodCount = 1
-	podsEvicted = removePodsWithAffinityRules(fakeClient, "v1", []*v1.Node{node}, false, npe, 1, false)
-	if podsEvicted != expectedEvictedPodCount {
-		t.Errorf("Unexpected no of pods evicted: pods evicted: %d, expected: %d", podsEvicted, expectedEvictedPodCount)
+
+	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: []v1.Pod{*p1, *p2, *p3, *p4}}, nil
+		})
+		fakeClient.Fake.AddReactor("get", "nodes", func(action core.Action) (bool, runtime.Object, error) {
+			return true, node, nil
+		})
+
+		podEvictor := evictions.NewPodEvictor(
+			fakeClient,
+			"v1",
+			false,
+			test.maxPodsToEvict,
+			[]*v1.Node{node},
+		)
+
+		RemovePodsViolatingInterPodAntiAffinity(ctx, fakeClient, api.DeschedulerStrategy{}, []*v1.Node{node}, false, podEvictor)
+		podsEvicted := podEvictor.TotalEvicted()
+		if podsEvicted != test.expectedEvictedPodCount {
+			t.Errorf("Unexpected no of pods evicted: pods evicted: %d, expected: %d", podsEvicted, test.expectedEvictedPodCount)
+		}
 	}
 }
 

pkg/descheduler/strategies/pod_lifetime.go

@@ -0,0 +1,71 @@
+/*
+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"
+
+	v1 "k8s.io/api/core/v1"
+	v1meta "k8s.io/apimachinery/pkg/apis/meta/v1"
+	clientset "k8s.io/client-go/kubernetes"
+	"k8s.io/klog"
+
+	"sigs.k8s.io/descheduler/pkg/api"
+	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
+	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
+)
+
+// 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, evictLocalStoragePods bool, podEvictor *evictions.PodEvictor) {
+	if strategy.Params.MaxPodLifeTimeSeconds == nil {
+		klog.V(1).Infof("MaxPodLifeTimeSeconds not set")
+		return
+	}
+
+	for _, node := range nodes {
+		klog.V(1).Infof("Processing node: %#v", node.Name)
+		pods := listOldPodsOnNode(ctx, client, node, *strategy.Params.MaxPodLifeTimeSeconds, evictLocalStoragePods)
+		for _, pod := range pods {
+			success, err := podEvictor.EvictPod(ctx, pod, node)
+			if success {
+				klog.V(1).Infof("Evicted pod: %#v because it was created more than %v seconds ago", pod.Name, *strategy.Params.MaxPodLifeTimeSeconds)
+			}
+
+			if err != nil {
+				klog.Errorf("Error evicting pod: (%#v)", err)
+				break
+			}
+		}
+	}
+}
+
+func listOldPodsOnNode(ctx context.Context, client clientset.Interface, node *v1.Node, maxAge uint, evictLocalStoragePods bool) []*v1.Pod {
+	pods, err := podutil.ListEvictablePodsOnNode(ctx, client, node, evictLocalStoragePods)
+	if err != nil {
+		return nil
+	}
+
+	var oldPods []*v1.Pod
+	for _, pod := range pods {
+		podAgeSeconds := uint(v1meta.Now().Sub(pod.GetCreationTimestamp().Local()).Seconds())
+		if podAgeSeconds > maxAge {
+			oldPods = append(oldPods, pod)
+		}
+	}
+
+	return oldPods
+}

pkg/descheduler/strategies/pod_lifetime_test.go

@@ -0,0 +1,169 @@
+/*
+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"
+	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()
+	node := 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, node.Name, nil)
+	p1.Namespace = "dev"
+	p1.ObjectMeta.CreationTimestamp = newerPodCreationTime
+	p2 := test.BuildTestPod("p2", 100, 0, node.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, node.Name, nil)
+	p3.Namespace = "dev"
+	p3.ObjectMeta.CreationTimestamp = newerPodCreationTime
+	p4 := test.BuildTestPod("p4", 100, 0, node.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, node.Name, nil)
+	p5.Namespace = "dev"
+	p5.ObjectMeta.CreationTimestamp = newerPodCreationTime
+	p6 := test.BuildTestPod("p6", 100, 0, node.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, node.Name, nil)
+	p7.Namespace = "dev"
+	p7.ObjectMeta.CreationTimestamp = newerPodCreationTime
+	p8 := test.BuildTestPod("p8", 100, 0, node.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
+
+	var maxLifeTime uint = 600
+	testCases := []struct {
+		description             string
+		strategy                api.DeschedulerStrategy
+		maxPodsToEvict          int
+		pods                    []v1.Pod
+		expectedEvictedPodCount int
+	}{
+		{
+			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{
+					MaxPodLifeTimeSeconds: &maxLifeTime,
+				},
+			},
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p1, *p2},
+			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{
+					MaxPodLifeTimeSeconds: &maxLifeTime,
+				},
+			},
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p3, *p4},
+			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{
+					MaxPodLifeTimeSeconds: &maxLifeTime,
+				},
+			},
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p5, *p6},
+			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{
+					MaxPodLifeTimeSeconds: &maxLifeTime,
+				},
+			},
+			maxPodsToEvict:          5,
+			pods:                    []v1.Pod{*p7, *p8},
+			expectedEvictedPodCount: 0,
+		},
+	}
+
+	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
+		})
+		fakeClient.Fake.AddReactor("get", "nodes", func(action core.Action) (bool, runtime.Object, error) {
+			return true, node, nil
+		})
+		podEvictor := evictions.NewPodEvictor(
+			fakeClient,
+			"v1",
+			false,
+			tc.maxPodsToEvict,
+			[]*v1.Node{node},
+		)
+
+		PodLifeTime(ctx, fakeClient, tc.strategy, []*v1.Node{node}, false, 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

@@ -0,0 +1,77 @@
+/*
+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"
+
+	v1 "k8s.io/api/core/v1"
+	clientset "k8s.io/client-go/kubernetes"
+	"k8s.io/klog"
+
+	"sigs.k8s.io/descheduler/pkg/api"
+	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
+	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
+)
+
+// 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, evictLocalStoragePods bool, podEvictor *evictions.PodEvictor) {
+	if strategy.Params.PodsHavingTooManyRestarts == nil || strategy.Params.PodsHavingTooManyRestarts.PodRestartThreshold < 1 {
+		klog.V(1).Infof("PodsHavingTooManyRestarts thresholds not set")
+		return
+	}
+	for _, node := range nodes {
+		klog.V(1).Infof("Processing node: %s", node.Name)
+		pods, err := podutil.ListEvictablePodsOnNode(ctx, client, node, evictLocalStoragePods)
+		if err != nil {
+			klog.Errorf("Error when list pods at node %s", node.Name)
+			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); err != nil {
+				klog.Errorf("Error evicting pod: (%#v)", err)
+				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

@@ -0,0 +1,183 @@
+/*
+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"
+
+	"k8s.io/api/core/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()
+	createStrategy := func(enabled, includingInitContainers bool, restartThresholds int32) api.DeschedulerStrategy {
+		return api.DeschedulerStrategy{
+			Enabled: enabled,
+			Params: api.StrategyParameters{
+				PodsHavingTooManyRestarts: &api.PodsHavingTooManyRestarts{
+					PodRestartThreshold:     restartThresholds,
+					IncludingInitContainers: includingInitContainers,
+				},
+			},
+		}
+	}
+
+	tests := []struct {
+		description             string
+		pods                    []v1.Pod
+		strategy                api.DeschedulerStrategy
+		expectedEvictedPodCount int
+		maxPodsToEvict          int
+	}{
+		{
+			description:             "All pods have total restarts under threshold, no pod evictions",
+			strategy:                createStrategy(true, true, 10000),
+			expectedEvictedPodCount: 0,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "Some pods have total restarts bigger than threshold",
+			strategy:                createStrategy(true, true, 1),
+			expectedEvictedPodCount: 6,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "Nine pods have total restarts equals threshold(includingInitContainers=true), 6 pods evictions",
+			strategy:                createStrategy(true, true, 1*25),
+			expectedEvictedPodCount: 6,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "Nine pods have total restarts equals threshold(includingInitContainers=false), 5 pods evictions",
+			strategy:                createStrategy(true, false, 1*25),
+			expectedEvictedPodCount: 5,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "All pods have total restarts equals threshold(includingInitContainers=true), 6 pods evictions",
+			strategy:                createStrategy(true, true, 1*20),
+			expectedEvictedPodCount: 6,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "Nine pods have total restarts equals threshold(includingInitContainers=false), 6 pods evictions",
+			strategy:                createStrategy(true, false, 1*20),
+			expectedEvictedPodCount: 6,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "Five pods have total restarts bigger than threshold(includingInitContainers=true), but only 1 pod eviction",
+			strategy:                createStrategy(true, true, 5*25+1),
+			expectedEvictedPodCount: 1,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "Five pods have total restarts bigger than threshold(includingInitContainers=false), but only 1 pod eviction",
+			strategy:                createStrategy(true, false, 5*20+1),
+			expectedEvictedPodCount: 1,
+			maxPodsToEvict:          0,
+		},
+		{
+			description:             "All pods have total restarts equals threshold(maxPodsToEvict=3), 3 pods evictions",
+			strategy:                createStrategy(true, true, 1),
+			expectedEvictedPodCount: 3,
+			maxPodsToEvict:          3,
+		},
+	}
+
+	for _, tc := range tests {
+		node := test.BuildTestNode("node1", 2000, 3000, 10, nil)
+		pods := initPods(node)
+
+		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,
+			"v1",
+			false,
+			tc.maxPodsToEvict,
+			[]*v1.Node{node},
+		)
+
+		RemovePodsHavingTooManyRestarts(ctx, fakeClient, tc.strategy, []*v1.Node{node}, false, 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/utils/pod.go

@@ -2,10 +2,12 @@ package utils
 
 import (
 	"fmt"
+
 	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/resource"
 	utilfeature "k8s.io/apiserver/pkg/util/feature"
 	"k8s.io/component-base/featuregate"
+	"k8s.io/klog"
 )
 
 const (
@@ -179,3 +181,15 @@ func maxResourceList(list, new v1.ResourceList) {
 		}
 	}
 }
+
+// PodToleratesTaints returns true if a pod tolerates one node's taints
+func PodToleratesTaints(pod *v1.Pod, taintsOfNodes map[string][]v1.Taint) bool {
+	for nodeName, taintsForNode := range taintsOfNodes {
+		if len(pod.Spec.Tolerations) >= len(taintsForNode) && TolerationsTolerateTaintsWithFilter(pod.Spec.Tolerations, taintsForNode, nil) {
+			return true
+		}
+		klog.V(5).Infof("pod: %#v doesn't tolerate node %s's taints", pod.Name, nodeName)
+	}
+
+	return false
+}

pkg/utils/predicates.go

@@ -19,7 +19,7 @@ package utils
 import (
 	"fmt"
 
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/labels"
 	"k8s.io/apimachinery/pkg/selection"
 	"k8s.io/klog"
@@ -126,3 +126,35 @@ func NodeSelectorRequirementsAsSelector(nsm []v1.NodeSelectorRequirement) (label
 	}
 	return selector, nil
 }
+
+// TolerationsTolerateTaint checks if taint is tolerated by any of the tolerations.
+func TolerationsTolerateTaint(tolerations []v1.Toleration, taint *v1.Taint) bool {
+	for i := range tolerations {
+		if tolerations[i].ToleratesTaint(taint) {
+			return true
+		}
+	}
+	return false
+}
+
+type taintsFilterFunc func(*v1.Taint) bool
+
+// TolerationsTolerateTaintsWithFilter checks if given tolerations tolerates
+// all the taints that apply to the filter in given taint list.
+func TolerationsTolerateTaintsWithFilter(tolerations []v1.Toleration, taints []v1.Taint, applyFilter taintsFilterFunc) bool {
+	if len(taints) == 0 {
+		return true
+	}
+
+	for i := range taints {
+		if applyFilter != nil && !applyFilter(&taints[i]) {
+			continue
+		}
+
+		if !TolerationsTolerateTaint(tolerations, &taints[i]) {
+			return false
+		}
+	}
+
+	return true
+}

pkg/utils/utils.go

@@ -1,35 +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 utils
-
-import v1 "k8s.io/api/core/v1"
-
-// This file contains the datastructures, types & functions needed by all the strategies so that we don't have
-// to compute them again in each strategy.
-
-// NodePodEvictedCount keeps count of pods evicted on node. This is used in conjunction with strategies to
-type NodePodEvictedCount map[*v1.Node]int
-
-// InitializeNodePodCount initializes the nodePodCount.
-func InitializeNodePodCount(nodeList []*v1.Node) NodePodEvictedCount {
-	var nodePodCount = make(NodePodEvictedCount)
-	for _, node := range nodeList {
-		// Initialize podsEvicted till now with 0.
-		nodePodCount[node] = 0
-	}
-	return nodePodCount
-}

test/e2e/e2e_test.go

@@ -17,6 +17,7 @@ limitations under the License.
 package e2e
 
 import (
+	"context"
 	"math"
 	"testing"
 	"time"
@@ -24,6 +25,8 @@ import (
 	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/client-go/informers"
+	coreinformers "k8s.io/client-go/informers/core/v1"
 	clientset "k8s.io/client-go/kubernetes"
 	"k8s.io/klog"
 	"sigs.k8s.io/descheduler/cmd/descheduler/app/options"
@@ -31,11 +34,11 @@ import (
 	deschedulerapi "sigs.k8s.io/descheduler/pkg/api"
 	"sigs.k8s.io/descheduler/pkg/descheduler"
 	"sigs.k8s.io/descheduler/pkg/descheduler/client"
+	"sigs.k8s.io/descheduler/pkg/descheduler/evictions"
 	eutils "sigs.k8s.io/descheduler/pkg/descheduler/evictions/utils"
 	nodeutil "sigs.k8s.io/descheduler/pkg/descheduler/node"
 	podutil "sigs.k8s.io/descheduler/pkg/descheduler/pod"
 	"sigs.k8s.io/descheduler/pkg/descheduler/strategies"
-	"sigs.k8s.io/descheduler/pkg/utils"
 )
 
 func MakePodSpec() v1.PodSpec {
@@ -93,7 +96,7 @@ func RcByNameContainer(name string, replicas int32, labels map[string]string, gr
 }
 
 // startEndToEndForLowNodeUtilization tests the lownode utilization strategy.
-func startEndToEndForLowNodeUtilization(clientset clientset.Interface) {
+func startEndToEndForLowNodeUtilization(ctx context.Context, clientset clientset.Interface, nodeInformer coreinformers.NodeInformer) {
 	var thresholds = make(deschedulerapi.ResourceThresholds)
 	var targetThresholds = make(deschedulerapi.ResourceThresholds)
 	thresholds[v1.ResourceMemory] = 20
@@ -108,38 +111,61 @@ func startEndToEndForLowNodeUtilization(clientset clientset.Interface) {
 		klog.Fatalf("%v", err)
 	}
 	stopChannel := make(chan struct{})
-	nodes, err := nodeutil.ReadyNodes(clientset, "", stopChannel)
+	nodes, err := nodeutil.ReadyNodes(ctx, clientset, nodeInformer, "", stopChannel)
 	if err != nil {
 		klog.Fatalf("%v", err)
 	}
-	nodeUtilizationThresholds := deschedulerapi.NodeResourceUtilizationThresholds{Thresholds: thresholds, TargetThresholds: targetThresholds}
-	nodeUtilizationStrategyParams := deschedulerapi.StrategyParameters{NodeResourceUtilizationThresholds: nodeUtilizationThresholds}
-	lowNodeUtilizationStrategy := deschedulerapi.DeschedulerStrategy{Enabled: true, Params: nodeUtilizationStrategyParams}
-	ds := &options.DeschedulerServer{Client: clientset}
-	nodePodCount := utils.InitializeNodePodCount(nodes)
-	strategies.LowNodeUtilization(ds, lowNodeUtilizationStrategy, evictionPolicyGroupVersion, nodes, nodePodCount)
+
+	lowNodeUtilizationStrategy := deschedulerapi.DeschedulerStrategy{
+		Enabled: true,
+		Params: deschedulerapi.StrategyParameters{
+			NodeResourceUtilizationThresholds: &deschedulerapi.NodeResourceUtilizationThresholds{
+				Thresholds:       thresholds,
+				TargetThresholds: targetThresholds,
+			},
+		},
+	}
+
+	podEvictor := evictions.NewPodEvictor(
+		clientset,
+		evictionPolicyGroupVersion,
+		false,
+		0,
+		nodes,
+	)
+
+	strategies.LowNodeUtilization(ctx, clientset, lowNodeUtilizationStrategy, nodes, false, podEvictor)
 	time.Sleep(10 * time.Second)
 }
 
 func TestE2E(t *testing.T) {
 	// If we have reached here, it means cluster would have been already setup and the kubeconfig file should
 	// be in /tmp directory as admin.conf.
+	ctx := context.Background()
 	clientSet, err := client.CreateClient("/tmp/admin.conf")
 	if err != nil {
 		t.Errorf("Error during client creation with %v", err)
 	}
-	nodeList, err := clientSet.CoreV1().Nodes().List(metav1.ListOptions{})
+	nodeList, err := clientSet.CoreV1().Nodes().List(ctx, metav1.ListOptions{})
 	if err != nil {
 		t.Errorf("Error listing node with %v", err)
 	}
+	sharedInformerFactory := informers.NewSharedInformerFactory(clientSet, 0)
+	nodeInformer := sharedInformerFactory.Core().V1().Nodes()
+
+	stopChannel := make(chan struct{}, 0)
+	sharedInformerFactory.Start(stopChannel)
+	sharedInformerFactory.WaitForCacheSync(stopChannel)
+	defer close(stopChannel)
+
 	// Assumption: We would have 3 node cluster by now. Kubeadm brings all the master components onto master node.
 	// So, the last node would have least utilization.
 	rc := RcByNameContainer("test-rc", int32(15), map[string]string{"test": "app"}, nil)
-	_, err = clientSet.CoreV1().ReplicationControllers("default").Create(rc)
+	_, err = clientSet.CoreV1().ReplicationControllers("default").Create(ctx, rc, metav1.CreateOptions{})
 	if err != nil {
 		t.Errorf("Error creating deployment %v", err)
 	}
-	evictPods(t, clientSet, nodeList, rc)
+	evictPods(ctx, t, clientSet, nodeInformer, nodeList, rc)
 
 	rc.Spec.Template.Annotations = map[string]string{"descheduler.alpha.kubernetes.io/evict": "true"}
 	rc.Spec.Replicas = func(i int32) *int32 { return &i }(15)
@@ -152,14 +178,15 @@ func TestE2E(t *testing.T) {
 			},
 		},
 	}
-	_, err = clientSet.CoreV1().ReplicationControllers("default").Create(rc)
+	_, err = clientSet.CoreV1().ReplicationControllers("default").Create(ctx, rc, metav1.CreateOptions{})
 	if err != nil {
 		t.Errorf("Error creating deployment %v", err)
 	}
-	evictPods(t, clientSet, nodeList, rc)
+	evictPods(ctx, t, clientSet, nodeInformer, nodeList, rc)
 }
 
 func TestDeschedulingInterval(t *testing.T) {
+	ctx := context.Background()
 	clientSet, err := client.CreateClient("/tmp/admin.conf")
 	if err != nil {
 		t.Errorf("Error during client creation with %v", err)
@@ -173,8 +200,13 @@ func TestDeschedulingInterval(t *testing.T) {
 
 	c := make(chan bool)
 	go func() {
-		err := descheduler.RunDeschedulerStrategies(s, deschedulerPolicy)
-		if err != nil {
+		evictionPolicyGroupVersion, err := eutils.SupportEviction(s.Client)
+		if err != nil || len(evictionPolicyGroupVersion) == 0 {
+			t.Errorf("Error when checking support for eviction: %v", err)
+		}
+
+		stopChannel := make(chan struct{})
+		if err := descheduler.RunDeschedulerStrategies(ctx, s, deschedulerPolicy, evictionPolicyGroupVersion, stopChannel); err != nil {
 			t.Errorf("Error running descheduler strategies: %+v", err)
 		}
 		c <- true
@@ -188,7 +220,7 @@ func TestDeschedulingInterval(t *testing.T) {
 	}
 }
 
-func evictPods(t *testing.T, clientSet clientset.Interface, nodeList *v1.NodeList, rc *v1.ReplicationController) {
+func evictPods(ctx context.Context, t *testing.T, clientSet clientset.Interface, nodeInformer coreinformers.NodeInformer, nodeList *v1.NodeList, rc *v1.ReplicationController) {
 	var leastLoadedNode v1.Node
 	podsBefore := math.MaxInt16
 	for i := range nodeList.Items {
@@ -197,7 +229,7 @@ func evictPods(t *testing.T, clientSet clientset.Interface, nodeList *v1.NodeLis
 			continue
 		}
 		// List all the pods on the current Node
-		podsOnANode, err := podutil.ListEvictablePodsOnNode(clientSet, &nodeList.Items[i], true)
+		podsOnANode, err := podutil.ListEvictablePodsOnNode(ctx, clientSet, &nodeList.Items[i], true)
 		if err != nil {
 			t.Errorf("Error listing pods on a node %v", err)
 		}
@@ -208,8 +240,8 @@ func evictPods(t *testing.T, clientSet clientset.Interface, nodeList *v1.NodeLis
 		}
 	}
 	t.Log("Eviction of pods starting")
-	startEndToEndForLowNodeUtilization(clientSet)
-	podsOnleastUtilizedNode, err := podutil.ListEvictablePodsOnNode(clientSet, &leastLoadedNode, true)
+	startEndToEndForLowNodeUtilization(ctx, clientSet, nodeInformer)
+	podsOnleastUtilizedNode, err := podutil.ListEvictablePodsOnNode(ctx, clientSet, &leastLoadedNode, true)
 	if err != nil {
 		t.Errorf("Error listing pods on a node %v", err)
 	}
@@ -220,14 +252,14 @@ func evictPods(t *testing.T, clientSet clientset.Interface, nodeList *v1.NodeLis
 
 	//set number of replicas to 0
 	rc.Spec.Replicas = func(i int32) *int32 { return &i }(0)
-	_, err = clientSet.CoreV1().ReplicationControllers("default").Update(rc)
+	_, err = clientSet.CoreV1().ReplicationControllers("default").Update(ctx, rc, metav1.UpdateOptions{})
 	if err != nil {
 		t.Errorf("Error updating replica controller %v", err)
 	}
 	allPodsDeleted := false
 	//wait 30 seconds until all pods are deleted
 	for i := 0; i < 6; i++ {
-		scale, _ := clientSet.CoreV1().ReplicationControllers("default").GetScale(rc.Name, metav1.GetOptions{})
+		scale, _ := clientSet.CoreV1().ReplicationControllers("default").GetScale(ctx, rc.Name, metav1.GetOptions{})
 		if scale.Spec.Replicas == 0 {
 			allPodsDeleted = true
 			break
@@ -239,7 +271,7 @@ func evictPods(t *testing.T, clientSet clientset.Interface, nodeList *v1.NodeLis
 		t.Errorf("Deleting of rc pods took too long")
 	}
 
-	err = clientSet.CoreV1().ReplicationControllers("default").Delete(rc.Name, &metav1.DeleteOptions{})
+	err = clientSet.CoreV1().ReplicationControllers("default").Delete(ctx, rc.Name, metav1.DeleteOptions{})
 	if err != nil {
 		t.Errorf("Error deleting rc %v", err)
 	}

test/run-e2e-tests.sh

@@ -15,5 +15,22 @@
 #!/bin/bash
 
 # This just run e2e tests.
+if [ -n "$KIND_E2E" ]; then
+    K8S_VERSION=${KUBERNETES_VERSION:-v1.18.2}
+    curl -Lo kubectl https://storage.googleapis.com/kubernetes-release/release/${K8S_VERSION}/bin/linux/amd64/kubectl && chmod +x kubectl && sudo mv kubectl /usr/local/bin/
+    wget https://github.com/kubernetes-sigs/kind/releases/download/v0.8.1/kind-linux-amd64
+    chmod +x kind-linux-amd64
+    mv kind-linux-amd64 kind
+    export PATH=$PATH:$PWD
+    kind create cluster --image kindest/node:${K8S_VERSION} --config=./hack/kind_config.yaml
+    export KUBECONFIG="$(kind get kubeconfig-path)"
+    docker pull kubernetes/pause
+    kind load docker-image kubernetes/pause
+    kind get kubeconfig > /tmp/admin.conf
+    export KUBECONFIG="/tmp/admin.conf"
+    mkdir -p ~/gopath/src/sigs.k8s.io/
+    mv ~/gopath/src/github.com/kubernetes-sigs/descheduler ~/gopath/src/sigs.k8s.io/.
+fi
+
 PRJ_PREFIX="sigs.k8s.io/descheduler"
 go test ${PRJ_PREFIX}/test/e2e/ -v

test/test_utils.go

@@ -25,7 +25,7 @@ import (
 )
 
 // BuildTestPod creates a test pod with given parameters.
-func BuildTestPod(name string, cpu int64, memory int64, nodeName string) *v1.Pod {
+func BuildTestPod(name string, cpu int64, memory int64, nodeName string, apply func(*v1.Pod)) *v1.Pod {
 	pod := &v1.Pod{
 		ObjectMeta: metav1.ObjectMeta{
 			Namespace: "default",
@@ -50,7 +50,9 @@ func BuildTestPod(name string, cpu int64, memory int64, nodeName string) *v1.Pod
 	if memory >= 0 {
 		pod.Spec.Containers[0].Resources.Requests[v1.ResourceMemory] = *resource.NewQuantity(memory, resource.DecimalSI)
 	}
-
+	if apply != nil {
+		apply(pod)
+	}
 	return pod
 }
 
@@ -85,7 +87,7 @@ func GetDaemonSetOwnerRefList() []metav1.OwnerReference {
 }
 
 // BuildTestNode creates a node with specified capacity.
-func BuildTestNode(name string, millicpu int64, mem int64, pods int64) *v1.Node {
+func BuildTestNode(name string, millicpu int64, mem int64, pods int64, apply func(*v1.Node)) *v1.Node {
 	node := &v1.Node{
 		ObjectMeta: metav1.ObjectMeta{
 			Name:     name,
@@ -109,5 +111,8 @@ func BuildTestNode(name string, millicpu int64, mem int64, pods int64) *v1.Node
 			},
 		},
 	}
+	if apply != nil {
+		apply(node)
+	}
 	return node
 }

vendor/github.com/gogo/protobuf/proto/encode.go

@@ -189,6 +189,8 @@ type Marshaler interface {
 // prefixed by a varint-encoded length.
 func (p *Buffer) EncodeMessage(pb Message) error {
 	siz := Size(pb)
+	sizVar := SizeVarint(uint64(siz))
+	p.grow(siz + sizVar)
 	p.EncodeVarint(uint64(siz))
 	return p.Marshal(pb)
 }

vendor/github.com/gogo/protobuf/proto/lib.go

@@ -948,13 +948,19 @@ func isProto3Zero(v reflect.Value) bool {
 	return false
 }
 
-// ProtoPackageIsVersion2 is referenced from generated protocol buffer files
-// to assert that that code is compatible with this version of the proto package.
-const GoGoProtoPackageIsVersion2 = true
+const (
+	// ProtoPackageIsVersion3 is referenced from generated protocol buffer files
+	// to assert that that code is compatible with this version of the proto package.
+	GoGoProtoPackageIsVersion3 = true
 
-// ProtoPackageIsVersion1 is referenced from generated protocol buffer files
-// to assert that that code is compatible with this version of the proto package.
-const GoGoProtoPackageIsVersion1 = true
+	// ProtoPackageIsVersion2 is referenced from generated protocol buffer files
+	// to assert that that code is compatible with this version of the proto package.
+	GoGoProtoPackageIsVersion2 = true
+
+	// ProtoPackageIsVersion1 is referenced from generated protocol buffer files
+	// to assert that that code is compatible with this version of the proto package.
+	GoGoProtoPackageIsVersion1 = true
+)
 
 // InternalMessageInfo is a type used internally by generated .pb.go files.
 // This type is not intended to be used by non-generated code.

vendor/github.com/gogo/protobuf/proto/properties.go

@@ -43,7 +43,6 @@ package proto
 import (
 	"fmt"
 	"log"
-	"os"
 	"reflect"
 	"sort"
 	"strconv"
@@ -205,7 +204,7 @@ func (p *Properties) Parse(s string) {
 	// "bytes,49,opt,name=foo,def=hello!"
 	fields := strings.Split(s, ",") // breaks def=, but handled below.
 	if len(fields) < 2 {
-		fmt.Fprintf(os.Stderr, "proto: tag has too few fields: %q\n", s)
+		log.Printf("proto: tag has too few fields: %q", s)
 		return
 	}
 
@@ -225,7 +224,7 @@ func (p *Properties) Parse(s string) {
 		p.WireType = WireBytes
 		// no numeric converter for non-numeric types
 	default:
-		fmt.Fprintf(os.Stderr, "proto: tag has unknown wire type: %q\n", s)
+		log.Printf("proto: tag has unknown wire type: %q", s)
 		return
 	}
 
@@ -400,6 +399,15 @@ func GetProperties(t reflect.Type) *StructProperties {
 	return sprop
 }
 
+type (
+	oneofFuncsIface interface {
+		XXX_OneofFuncs() (func(Message, *Buffer) error, func(Message, int, int, *Buffer) (bool, error), func(Message) int, []interface{})
+	}
+	oneofWrappersIface interface {
+		XXX_OneofWrappers() []interface{}
+	}
+)
+
 // getPropertiesLocked requires that propertiesMu is held.
 func getPropertiesLocked(t reflect.Type) *StructProperties {
 	if prop, ok := propertiesMap[t]; ok {
@@ -441,37 +449,40 @@ func getPropertiesLocked(t reflect.Type) *StructProperties {
 	// Re-order prop.order.
 	sort.Sort(prop)
 
-	type oneofMessage interface {
-		XXX_OneofFuncs() (func(Message, *Buffer) error, func(Message, int, int, *Buffer) (bool, error), func(Message) int, []interface{})
-	}
-	if om, ok := reflect.Zero(reflect.PtrTo(t)).Interface().(oneofMessage); isOneofMessage && ok {
+	if isOneofMessage {
 		var oots []interface{}
-		_, _, _, oots = om.XXX_OneofFuncs()
-
-		// Interpret oneof metadata.
-		prop.OneofTypes = make(map[string]*OneofProperties)
-		for _, oot := range oots {
-			oop := &OneofProperties{
-				Type: reflect.ValueOf(oot).Type(), // *T
-				Prop: new(Properties),
-			}
-			sft := oop.Type.Elem().Field(0)
-			oop.Prop.Name = sft.Name
-			oop.Prop.Parse(sft.Tag.Get("protobuf"))
-			// There will be exactly one interface field that
-			// this new value is assignable to.
-			for i := 0; i < t.NumField(); i++ {
-				f := t.Field(i)
-				if f.Type.Kind() != reflect.Interface {
-					continue
+		switch m := reflect.Zero(reflect.PtrTo(t)).Interface().(type) {
+		case oneofFuncsIface:
+			_, _, _, oots = m.XXX_OneofFuncs()
+		case oneofWrappersIface:
+			oots = m.XXX_OneofWrappers()
+		}
+		if len(oots) > 0 {
+			// Interpret oneof metadata.
+			prop.OneofTypes = make(map[string]*OneofProperties)
+			for _, oot := range oots {
+				oop := &OneofProperties{
+					Type: reflect.ValueOf(oot).Type(), // *T
+					Prop: new(Properties),
 				}
-				if !oop.Type.AssignableTo(f.Type) {
-					continue
+				sft := oop.Type.Elem().Field(0)
+				oop.Prop.Name = sft.Name
+				oop.Prop.Parse(sft.Tag.Get("protobuf"))
+				// There will be exactly one interface field that
+				// this new value is assignable to.
+				for i := 0; i < t.NumField(); i++ {
+					f := t.Field(i)
+					if f.Type.Kind() != reflect.Interface {
+						continue
+					}
+					if !oop.Type.AssignableTo(f.Type) {
+						continue
+					}
+					oop.Field = i
+					break
 				}
-				oop.Field = i
-				break
+				prop.OneofTypes[oop.Prop.OrigName] = oop
 			}
-			prop.OneofTypes[oop.Prop.OrigName] = oop
 		}
 	}
 

vendor/github.com/gogo/protobuf/proto/table_marshal.go

@@ -389,8 +389,13 @@ func (u *marshalInfo) computeMarshalInfo() {
 	// get oneof implementers
 	var oneofImplementers []interface{}
 	// gogo: isOneofMessage is needed for embedded oneof messages, without a marshaler and unmarshaler
-	if m, ok := reflect.Zero(reflect.PtrTo(t)).Interface().(oneofMessage); ok && isOneofMessage {
-		_, _, _, oneofImplementers = m.XXX_OneofFuncs()
+	if isOneofMessage {
+		switch m := reflect.Zero(reflect.PtrTo(t)).Interface().(type) {
+		case oneofFuncsIface:
+			_, _, _, oneofImplementers = m.XXX_OneofFuncs()
+		case oneofWrappersIface:
+			oneofImplementers = m.XXX_OneofWrappers()
+		}
 	}
 
 	// normal fields
@@ -519,10 +524,6 @@ func (fi *marshalFieldInfo) computeOneofFieldInfo(f *reflect.StructField, oneofI
 	}
 }
 
-type oneofMessage interface {
-	XXX_OneofFuncs() (func(Message, *Buffer) error, func(Message, int, int, *Buffer) (bool, error), func(Message) int, []interface{})
-}
-
 // wiretype returns the wire encoding of the type.
 func wiretype(encoding string) uint64 {
 	switch encoding {
@@ -2968,7 +2969,9 @@ func (p *Buffer) Marshal(pb Message) error {
 	if m, ok := pb.(newMarshaler); ok {
 		siz := m.XXX_Size()
 		p.grow(siz) // make sure buf has enough capacity
-		p.buf, err = m.XXX_Marshal(p.buf, p.deterministic)
+		pp := p.buf[len(p.buf) : len(p.buf) : len(p.buf)+siz]
+		pp, err = m.XXX_Marshal(pp, p.deterministic)
+		p.buf = append(p.buf, pp...)
 		return err
 	}
 	if m, ok := pb.(Marshaler); ok {

vendor/github.com/gogo/protobuf/proto/table_merge.go

@@ -530,6 +530,25 @@ func (mi *mergeInfo) computeMergeInfo() {
 			}
 		case reflect.Struct:
 			switch {
+			case isSlice && !isPointer: // E.g. []pb.T
+				mergeInfo := getMergeInfo(tf)
+				zero := reflect.Zero(tf)
+				mfi.merge = func(dst, src pointer) {
+					// TODO: Make this faster?
+					dstsp := dst.asPointerTo(f.Type)
+					dsts := dstsp.Elem()
+					srcs := src.asPointerTo(f.Type).Elem()
+					for i := 0; i < srcs.Len(); i++ {
+						dsts = reflect.Append(dsts, zero)
+						srcElement := srcs.Index(i).Addr()
+						dstElement := dsts.Index(dsts.Len() - 1).Addr()
+						mergeInfo.merge(valToPointer(dstElement), valToPointer(srcElement))
+					}
+					if dsts.IsNil() {
+						dsts = reflect.MakeSlice(f.Type, 0, 0)
+					}
+					dstsp.Elem().Set(dsts)
+				}
 			case !isPointer:
 				mergeInfo := getMergeInfo(tf)
 				mfi.merge = func(dst, src pointer) {

vendor/github.com/gogo/protobuf/proto/table_unmarshal.go

@@ -371,15 +371,18 @@ func (u *unmarshalInfo) computeUnmarshalInfo() {
 	}
 
 	// Find any types associated with oneof fields.
-	// TODO: XXX_OneofFuncs returns more info than we need.  Get rid of some of it?
-	fn := reflect.Zero(reflect.PtrTo(t)).MethodByName("XXX_OneofFuncs")
 	// gogo: len(oneofFields) > 0 is needed for embedded oneof messages, without a marshaler and unmarshaler
-	if fn.IsValid() && len(oneofFields) > 0 {
-		res := fn.Call(nil)[3] // last return value from XXX_OneofFuncs: []interface{}
-		for i := res.Len() - 1; i >= 0; i-- {
-			v := res.Index(i)                             // interface{}
-			tptr := reflect.ValueOf(v.Interface()).Type() // *Msg_X
-			typ := tptr.Elem()                            // Msg_X
+	if len(oneofFields) > 0 {
+		var oneofImplementers []interface{}
+		switch m := reflect.Zero(reflect.PtrTo(t)).Interface().(type) {
+		case oneofFuncsIface:
+			_, _, _, oneofImplementers = m.XXX_OneofFuncs()
+		case oneofWrappersIface:
+			oneofImplementers = m.XXX_OneofWrappers()
+		}
+		for _, v := range oneofImplementers {
+			tptr := reflect.TypeOf(v) // *Msg_X
+			typ := tptr.Elem()        // Msg_X
 
 			f := typ.Field(0) // oneof implementers have one field
 			baseUnmarshal := fieldUnmarshaler(&f)
@@ -407,11 +410,12 @@ func (u *unmarshalInfo) computeUnmarshalInfo() {
 					u.setTag(fieldNum, of.field, unmarshal, 0, name)
 				}
 			}
+
 		}
 	}
 
 	// Get extension ranges, if any.
-	fn = reflect.Zero(reflect.PtrTo(t)).MethodByName("ExtensionRangeArray")
+	fn := reflect.Zero(reflect.PtrTo(t)).MethodByName("ExtensionRangeArray")
 	if fn.IsValid() {
 		if !u.extensions.IsValid() && !u.oldExtensions.IsValid() && !u.bytesExtensions.IsValid() {
 			panic("a message with extensions, but no extensions field in " + t.Name())

vendor/github.com/gogo/protobuf/proto/text.go

@@ -476,6 +476,8 @@ func (tm *TextMarshaler) writeStruct(w *textWriter, sv reflect.Value) error {
 	return nil
 }
 
+var textMarshalerType = reflect.TypeOf((*encoding.TextMarshaler)(nil)).Elem()
+
 // writeAny writes an arbitrary field.
 func (tm *TextMarshaler) writeAny(w *textWriter, v reflect.Value, props *Properties) error {
 	v = reflect.Indirect(v)
@@ -589,8 +591,8 @@ func (tm *TextMarshaler) writeAny(w *textWriter, v reflect.Value, props *Propert
 			// mutating this value.
 			v = v.Addr()
 		}
-		if etm, ok := v.Interface().(encoding.TextMarshaler); ok {
-			text, err := etm.MarshalText()
+		if v.Type().Implements(textMarshalerType) {
+			text, err := v.Interface().(encoding.TextMarshaler).MarshalText()
 			if err != nil {
 				return err
 			}

vendor/github.com/google/gofuzz/README.md

@@ -3,7 +3,7 @@ gofuzz
 
 gofuzz is a library for populating go objects with random values.
 
-[![GoDoc](https://godoc.org/github.com/google/gofuzz?status.png)](https://godoc.org/github.com/google/gofuzz)
+[![GoDoc](https://godoc.org/github.com/google/gofuzz?status.svg)](https://godoc.org/github.com/google/gofuzz)
 [![Travis](https://travis-ci.org/google/gofuzz.svg?branch=master)](https://travis-ci.org/google/gofuzz)
 
 This is useful for testing:

vendor/github.com/google/gofuzz/fuzz.go

@@ -20,6 +20,7 @@ import (
 	"fmt"
 	"math/rand"
 	"reflect"
+	"regexp"
 	"time"
 )
 
@@ -28,13 +29,14 @@ type fuzzFuncMap map[reflect.Type]reflect.Value
 
 // Fuzzer knows how to fill any object with random fields.
 type Fuzzer struct {
-	fuzzFuncs        fuzzFuncMap
-	defaultFuzzFuncs fuzzFuncMap
-	r                *rand.Rand
-	nilChance        float64
-	minElements      int
-	maxElements      int
-	maxDepth         int
+	fuzzFuncs         fuzzFuncMap
+	defaultFuzzFuncs  fuzzFuncMap
+	r                 *rand.Rand
+	nilChance         float64
+	minElements       int
+	maxElements       int
+	maxDepth          int
+	skipFieldPatterns []*regexp.Regexp
 }
 
 // New returns a new Fuzzer. Customize your Fuzzer further by calling Funcs,
@@ -150,6 +152,13 @@ func (f *Fuzzer) MaxDepth(d int) *Fuzzer {
 	return f
 }
 
+// Skip fields which match the supplied pattern. Call this multiple times if needed
+// This is useful to skip XXX_ fields generated by protobuf
+func (f *Fuzzer) SkipFieldsWithPattern(pattern *regexp.Regexp) *Fuzzer {
+	f.skipFieldPatterns = append(f.skipFieldPatterns, pattern)
+	return f
+}
+
 // Fuzz recursively fills all of obj's fields with something random.  First
 // this tries to find a custom fuzz function (see Funcs).  If there is no
 // custom function this tests whether the object implements fuzz.Interface and,
@@ -274,7 +283,17 @@ func (fc *fuzzerContext) doFuzz(v reflect.Value, flags uint64) {
 		v.Set(reflect.Zero(v.Type()))
 	case reflect.Struct:
 		for i := 0; i < v.NumField(); i++ {
-			fc.doFuzz(v.Field(i), 0)
+			skipField := false
+			fieldName := v.Type().Field(i).Name
+			for _, pattern := range fc.fuzzer.skipFieldPatterns {
+				if pattern.MatchString(fieldName) {
+					skipField = true
+					break
+				}
+			}
+			if !skipField {
+				fc.doFuzz(v.Field(i), 0)
+			}
 		}
 	case reflect.Chan:
 		fallthrough

vendor/github.com/googleapis/gnostic/compiler/reader.go

@@ -15,6 +15,7 @@
 package compiler
 
 import (
+	"errors"
 	"fmt"
 	"gopkg.in/yaml.v2"
 	"io/ioutil"
@@ -53,11 +54,16 @@ func FetchFile(fileurl string) ([]byte, error) {
 		}
 		return bytes, nil
 	}
-	log.Printf("Fetching %s", fileurl)
+	if verboseReader {
+		log.Printf("Fetching %s", fileurl)
+	}
 	response, err := http.Get(fileurl)
 	if err != nil {
 		return nil, err
 	}
+	if response.StatusCode != 200 {
+		return nil, errors.New(fmt.Sprintf("Error downloading %s: %s", fileurl, response.Status))
+	}
 	defer response.Body.Close()
 	bytes, err = ioutil.ReadAll(response.Body)
 	if err == nil {

vendor/github.com/googleapis/gnostic/extensions/extension.pb.go

@@ -193,7 +193,7 @@ func init() { proto.RegisterFile("extension.proto", fileDescriptor0) }
 
 var fileDescriptor0 = []byte{
 	// 355 bytes of a gzipped FileDescriptorProto
-	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x09, 0x6e, 0x88, 0x02, 0xff, 0x74, 0x91, 0x4d, 0x4b, 0xf3, 0x40,
+	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0xff, 0x74, 0x91, 0x4d, 0x4b, 0xf3, 0x40,
 	0x1c, 0xc4, 0x49, 0xdf, 0xf2, 0x64, 0x1f, 0xb4, 0xb2, 0x16, 0x8d, 0xe2, 0xa1, 0x04, 0x84, 0x22,
 	0xb8, 0xa5, 0x0a, 0xde, 0x5b, 0x28, 0xea, 0xc5, 0x96, 0x3d, 0xd4, 0x9b, 0x65, 0x9b, 0xfe, 0xdb,
 	0x46, 0x92, 0xdd, 0x75, 0xf3, 0x62, 0xfb, 0x55, 0x3c, 0xfa, 0x49, 0x25, 0xbb, 0xd9, 0x7a, 0x50,

vendor/golang.org/x/crypto/ssh/terminal/terminal.go

@@ -7,6 +7,7 @@ package terminal
 import (
 	"bytes"
 	"io"
+	"runtime"
 	"strconv"
 	"sync"
 	"unicode/utf8"
@@ -939,6 +940,8 @@ func (s *stRingBuffer) NthPreviousEntry(n int) (value string, ok bool) {
 // readPasswordLine reads from reader until it finds \n or io.EOF.
 // The slice returned does not include the \n.
 // readPasswordLine also ignores any \r it finds.
+// Windows uses \r as end of line. So, on Windows, readPasswordLine
+// reads until it finds \r and ignores any \n it finds during processing.
 func readPasswordLine(reader io.Reader) ([]byte, error) {
 	var buf [1]byte
 	var ret []byte
@@ -947,10 +950,20 @@ func readPasswordLine(reader io.Reader) ([]byte, error) {
 		n, err := reader.Read(buf[:])
 		if n > 0 {
 			switch buf[0] {
+			case '\b':
+				if len(ret) > 0 {
+					ret = ret[:len(ret)-1]
+				}
 			case '\n':
-				return ret, nil
+				if runtime.GOOS != "windows" {
+					return ret, nil
+				}
+				// otherwise ignore \n
 			case '\r':
-				// remove \r from passwords on Windows
+				if runtime.GOOS == "windows" {
+					return ret, nil
+				}
+				// otherwise ignore \r
 			default:
 				ret = append(ret, buf[0])
 			}

vendor/golang.org/x/crypto/ssh/terminal/util_windows.go

@@ -85,8 +85,8 @@ func ReadPassword(fd int) ([]byte, error) {
 	}
 	old := st
 
-	st &^= (windows.ENABLE_ECHO_INPUT)
-	st |= (windows.ENABLE_PROCESSED_INPUT | windows.ENABLE_LINE_INPUT | windows.ENABLE_PROCESSED_OUTPUT)
+	st &^= (windows.ENABLE_ECHO_INPUT | windows.ENABLE_LINE_INPUT)
+	st |= (windows.ENABLE_PROCESSED_OUTPUT | windows.ENABLE_PROCESSED_INPUT)
 	if err := windows.SetConsoleMode(windows.Handle(fd), st); err != nil {
 		return nil, err
 	}

vendor/golang.org/x/sys/unix/affinity_linux.go

@@ -7,6 +7,7 @@
 package unix
 
 import (
+	"math/bits"
 	"unsafe"
 )
 
@@ -79,50 +80,7 @@ func (s *CPUSet) IsSet(cpu int) bool {
 func (s *CPUSet) Count() int {
 	c := 0
 	for _, b := range s {
-		c += onesCount64(uint64(b))
+		c += bits.OnesCount64(uint64(b))
 	}
 	return c
 }
-
-// onesCount64 is a copy of Go 1.9's math/bits.OnesCount64.
-// Once this package can require Go 1.9, we can delete this
-// and update the caller to use bits.OnesCount64.
-func onesCount64(x uint64) int {
-	const m0 = 0x5555555555555555 // 01010101 ...
-	const m1 = 0x3333333333333333 // 00110011 ...
-	const m2 = 0x0f0f0f0f0f0f0f0f // 00001111 ...
-
-	// Unused in this function, but definitions preserved for
-	// documentation purposes:
-	//
-	//   const m3 = 0x00ff00ff00ff00ff // etc.
-	//   const m4 = 0x0000ffff0000ffff
-	//
-	// Implementation: Parallel summing of adjacent bits.
-	// See "Hacker's Delight", Chap. 5: Counting Bits.
-	// The following pattern shows the general approach:
-	//
-	//   x = x>>1&(m0&m) + x&(m0&m)
-	//   x = x>>2&(m1&m) + x&(m1&m)
-	//   x = x>>4&(m2&m) + x&(m2&m)
-	//   x = x>>8&(m3&m) + x&(m3&m)
-	//   x = x>>16&(m4&m) + x&(m4&m)
-	//   x = x>>32&(m5&m) + x&(m5&m)
-	//   return int(x)
-	//
-	// Masking (& operations) can be left away when there's no
-	// danger that a field's sum will carry over into the next
-	// field: Since the result cannot be > 64, 8 bits is enough
-	// and we can ignore the masks for the shifts by 8 and up.
-	// Per "Hacker's Delight", the first line can be simplified
-	// more, but it saves at best one instruction, so we leave
-	// it alone for clarity.
-	const m = 1<<64 - 1
-	x = x>>1&(m0&m) + x&(m0&m)
-	x = x>>2&(m1&m) + x&(m1&m)
-	x = (x>>4 + x) & (m2 & m)
-	x += x >> 8
-	x += x >> 16
-	x += x >> 32
-	return int(x) & (1<<7 - 1)
-}

vendor/golang.org/x/sys/unix/ioctl.go

@@ -6,7 +6,19 @@
 
 package unix
 
-import "runtime"
+import (
+	"runtime"
+	"unsafe"
+)
+
+// ioctl itself should not be exposed directly, but additional get/set
+// functions for specific types are permissible.
+
+// IoctlSetInt performs an ioctl operation which sets an integer value
+// on fd, using the specified request number.
+func IoctlSetInt(fd int, req uint, value int) error {
+	return ioctl(fd, req, uintptr(value))
+}
 
 // IoctlSetWinsize performs an ioctl on fd with a *Winsize argument.
 //
@@ -14,7 +26,7 @@ import "runtime"
 func IoctlSetWinsize(fd int, req uint, value *Winsize) error {
 	// TODO: if we get the chance, remove the req parameter and
 	// hardcode TIOCSWINSZ.
-	err := ioctlSetWinsize(fd, req, value)
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(value)))
 	runtime.KeepAlive(value)
 	return err
 }
@@ -24,7 +36,30 @@ func IoctlSetWinsize(fd int, req uint, value *Winsize) error {
 // The req value will usually be TCSETA or TIOCSETA.
 func IoctlSetTermios(fd int, req uint, value *Termios) error {
 	// TODO: if we get the chance, remove the req parameter.
-	err := ioctlSetTermios(fd, req, value)
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(value)))
 	runtime.KeepAlive(value)
 	return err
 }
+
+// IoctlGetInt performs an ioctl operation which gets an integer value
+// from fd, using the specified request number.
+//
+// A few ioctl requests use the return value as an output parameter;
+// for those, IoctlRetInt should be used instead of this function.
+func IoctlGetInt(fd int, req uint) (int, error) {
+	var value int
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
+	return value, err
+}
+
+func IoctlGetWinsize(fd int, req uint) (*Winsize, error) {
+	var value Winsize
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
+	return &value, err
+}
+
+func IoctlGetTermios(fd int, req uint) (*Termios, error) {
+	var value Termios
+	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
+	return &value, err
+}

vendor/golang.org/x/sys/unix/mkall.sh

@@ -212,9 +212,11 @@ esac
 				echo "$mksyscall -tags $GOOS,$GOARCH $syscall_goos $GOOSARCH_in && gofmt -w zsyscall_$GOOSARCH.go && gofmt -w zsyscall_"$GOOSARCH"_gccgo.go && gofmt -w zsyscall_"$GOOSARCH"_gc.go " ;
 			elif [ "$GOOS" == "darwin" ]; then
 			        # pre-1.12, direct syscalls
-			        echo "$mksyscall -tags $GOOS,$GOARCH,!go1.12 $syscall_goos $GOOSARCH_in |gofmt >zsyscall_$GOOSARCH.1_11.go";
+			        echo "$mksyscall -tags $GOOS,$GOARCH,!go1.12 $syscall_goos syscall_darwin_${GOARCH}.1_11.go $GOOSARCH_in |gofmt >zsyscall_$GOOSARCH.1_11.go";
 			        # 1.12 and later, syscalls via libSystem
 				echo "$mksyscall -tags $GOOS,$GOARCH,go1.12 $syscall_goos $GOOSARCH_in |gofmt >zsyscall_$GOOSARCH.go";
+				# 1.13 and later, syscalls via libSystem (including syscallPtr)
+				echo "$mksyscall -tags $GOOS,$GOARCH,go1.13 syscall_darwin.1_13.go |gofmt >zsyscall_$GOOSARCH.1_13.go";
 			else
 				echo "$mksyscall -tags $GOOS,$GOARCH $syscall_goos $GOOSARCH_in |gofmt >zsyscall_$GOOSARCH.go";
 			fi

vendor/golang.org/x/sys/unix/mkerrors.sh

@@ -60,6 +60,7 @@ includes_Darwin='
 #include <sys/types.h>
 #include <sys/event.h>
 #include <sys/ptrace.h>
+#include <sys/select.h>
 #include <sys/socket.h>
 #include <sys/sockio.h>
 #include <sys/sysctl.h>
@@ -80,6 +81,7 @@ includes_Darwin='
 includes_DragonFly='
 #include <sys/types.h>
 #include <sys/event.h>
+#include <sys/select.h>
 #include <sys/socket.h>
 #include <sys/sockio.h>
 #include <sys/stat.h>
@@ -103,6 +105,7 @@ includes_FreeBSD='
 #include <sys/param.h>
 #include <sys/types.h>
 #include <sys/event.h>
+#include <sys/select.h>
 #include <sys/socket.h>
 #include <sys/sockio.h>
 #include <sys/stat.h>
@@ -179,24 +182,31 @@ struct ltchars {
 #include <sys/stat.h>
 #include <sys/types.h>
 #include <sys/time.h>
+#include <sys/select.h>
 #include <sys/signalfd.h>
 #include <sys/socket.h>
 #include <sys/xattr.h>
 #include <linux/bpf.h>
+#include <linux/can.h>
 #include <linux/capability.h>
+#include <linux/cryptouser.h>
 #include <linux/errqueue.h>
+#include <linux/falloc.h>
+#include <linux/fanotify.h>
+#include <linux/filter.h>
+#include <linux/fs.h>
+#include <linux/genetlink.h>
+#include <linux/hdreg.h>
+#include <linux/icmpv6.h>
 #include <linux/if.h>
+#include <linux/if_addr.h>
 #include <linux/if_alg.h>
 #include <linux/if_arp.h>
 #include <linux/if_ether.h>
 #include <linux/if_ppp.h>
 #include <linux/if_tun.h>
 #include <linux/if_packet.h>
-#include <linux/if_addr.h>
-#include <linux/falloc.h>
-#include <linux/fanotify.h>
-#include <linux/filter.h>
-#include <linux/fs.h>
+#include <linux/if_xdp.h>
 #include <linux/kexec.h>
 #include <linux/keyctl.h>
 #include <linux/loop.h>
@@ -206,26 +216,23 @@ struct ltchars {
 #include <linux/netfilter/nfnetlink.h>
 #include <linux/netlink.h>
 #include <linux/net_namespace.h>
+#include <linux/nsfs.h>
 #include <linux/perf_event.h>
+#include <linux/ptrace.h>
 #include <linux/random.h>
 #include <linux/reboot.h>
+#include <linux/rtc.h>
 #include <linux/rtnetlink.h>
-#include <linux/ptrace.h>
 #include <linux/sched.h>
 #include <linux/seccomp.h>
-#include <linux/sockios.h>
-#include <linux/wait.h>
-#include <linux/icmpv6.h>
 #include <linux/serial.h>
-#include <linux/can.h>
-#include <linux/vm_sockets.h>
+#include <linux/sockios.h>
 #include <linux/taskstats.h>
-#include <linux/genetlink.h>
+#include <linux/tipc.h>
+#include <linux/vm_sockets.h>
+#include <linux/wait.h>
 #include <linux/watchdog.h>
-#include <linux/hdreg.h>
-#include <linux/rtc.h>
-#include <linux/if_xdp.h>
-#include <linux/cryptouser.h>
+
 #include <mtd/ubi-user.h>
 #include <net/route.h>
 
@@ -264,6 +271,11 @@ struct ltchars {
 #define FS_KEY_DESC_PREFIX              "fscrypt:"
 #define FS_KEY_DESC_PREFIX_SIZE         8
 #define FS_MAX_KEY_SIZE                 64
+
+// The code generator produces -0x1 for (~0), but an unsigned value is necessary
+// for the tipc_subscr timeout __u32 field.
+#undef TIPC_WAIT_FOREVER
+#define TIPC_WAIT_FOREVER 0xffffffff
 '
 
 includes_NetBSD='
@@ -273,6 +285,7 @@ includes_NetBSD='
 #include <sys/extattr.h>
 #include <sys/mman.h>
 #include <sys/mount.h>
+#include <sys/select.h>
 #include <sys/socket.h>
 #include <sys/sockio.h>
 #include <sys/sysctl.h>
@@ -299,6 +312,7 @@ includes_OpenBSD='
 #include <sys/event.h>
 #include <sys/mman.h>
 #include <sys/mount.h>
+#include <sys/select.h>
 #include <sys/socket.h>
 #include <sys/sockio.h>
 #include <sys/stat.h>
@@ -335,6 +349,7 @@ includes_OpenBSD='
 includes_SunOS='
 #include <limits.h>
 #include <sys/types.h>
+#include <sys/select.h>
 #include <sys/socket.h>
 #include <sys/sockio.h>
 #include <sys/stat.h>
@@ -427,6 +442,7 @@ ccflags="$@"
 		$2 == "XCASE" ||
 		$2 == "ALTWERASE" ||
 		$2 == "NOKERNINFO" ||
+		$2 == "NFDBITS" ||
 		$2 ~ /^PAR/ ||
 		$2 ~ /^SIG[^_]/ ||
 		$2 ~ /^O[CNPFPL][A-Z]+[^_][A-Z]+$/ ||
@@ -451,6 +467,7 @@ ccflags="$@"
 		$2 ~ /^SYSCTL_VERS/ ||
 		$2 !~ "MNT_BITS" &&
 		$2 ~ /^(MS|MNT|UMOUNT)_/ ||
+		$2 ~ /^NS_GET_/ ||
 		$2 ~ /^TUN(SET|GET|ATTACH|DETACH)/ ||
 		$2 ~ /^(O|F|[ES]?FD|NAME|S|PTRACE|PT)_/ ||
 		$2 ~ /^KEXEC_/ ||
@@ -506,6 +523,7 @@ ccflags="$@"
 		$2 ~ /^XDP_/ ||
 		$2 ~ /^(HDIO|WIN|SMART)_/ ||
 		$2 ~ /^CRYPTO_/ ||
+		$2 ~ /^TIPC_/ ||
 		$2 !~ "WMESGLEN" &&
 		$2 ~ /^W[A-Z0-9]+$/ ||
 		$2 ~/^PPPIOC/ ||

vendor/golang.org/x/sys/unix/sockcmsg_dragonfly.go

@@ -0,0 +1,16 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package unix
+
+// Round the length of a raw sockaddr up to align it properly.
+func cmsgAlignOf(salen int) int {
+	salign := SizeofPtr
+	if SizeofPtr == 8 && !supportsABI(_dragonflyABIChangeVersion) {
+		// 64-bit Dragonfly before the September 2019 ABI changes still requires
+		// 32-bit aligned access to network subsystem.
+		salign = 4
+	}
+	return (salen + salign - 1) & ^(salign - 1)
+}

vendor/golang.org/x/sys/unix/sockcmsg_linux.go

@@ -17,7 +17,7 @@ func UnixCredentials(ucred *Ucred) []byte {
 	h.Level = SOL_SOCKET
 	h.Type = SCM_CREDENTIALS
 	h.SetLen(CmsgLen(SizeofUcred))
-	*((*Ucred)(cmsgData(h))) = *ucred
+	*(*Ucred)(h.data(0)) = *ucred
 	return b
 }
 

vendor/golang.org/x/sys/unix/sockcmsg_unix.go

@@ -9,35 +9,9 @@
 package unix
 
 import (
-	"runtime"
 	"unsafe"
 )
 
-// Round the length of a raw sockaddr up to align it properly.
-func cmsgAlignOf(salen int) int {
-	salign := SizeofPtr
-
-	switch runtime.GOOS {
-	case "aix":
-		// There is no alignment on AIX.
-		salign = 1
-	case "darwin", "dragonfly", "solaris", "illumos":
-		// NOTE: It seems like 64-bit Darwin, DragonFly BSD,
-		// illumos, and Solaris kernels still require 32-bit
-		// aligned access to network subsystem.
-		if SizeofPtr == 8 {
-			salign = 4
-		}
-	case "netbsd", "openbsd":
-		// NetBSD and OpenBSD armv7 require 64-bit alignment.
-		if runtime.GOARCH == "arm" {
-			salign = 8
-		}
-	}
-
-	return (salen + salign - 1) & ^(salign - 1)
-}
-
 // CmsgLen returns the value to store in the Len field of the Cmsghdr
 // structure, taking into account any necessary alignment.
 func CmsgLen(datalen int) int {
@@ -50,8 +24,8 @@ func CmsgSpace(datalen int) int {
 	return cmsgAlignOf(SizeofCmsghdr) + cmsgAlignOf(datalen)
 }
 
-func cmsgData(h *Cmsghdr) unsafe.Pointer {
-	return unsafe.Pointer(uintptr(unsafe.Pointer(h)) + uintptr(cmsgAlignOf(SizeofCmsghdr)))
+func (h *Cmsghdr) data(offset uintptr) unsafe.Pointer {
+	return unsafe.Pointer(uintptr(unsafe.Pointer(h)) + uintptr(cmsgAlignOf(SizeofCmsghdr)) + offset)
 }
 
 // SocketControlMessage represents a socket control message.
@@ -94,10 +68,8 @@ func UnixRights(fds ...int) []byte {
 	h.Level = SOL_SOCKET
 	h.Type = SCM_RIGHTS
 	h.SetLen(CmsgLen(datalen))
-	data := cmsgData(h)
-	for _, fd := range fds {
-		*(*int32)(data) = int32(fd)
-		data = unsafe.Pointer(uintptr(data) + 4)
+	for i, fd := range fds {
+		*(*int32)(h.data(4 * uintptr(i))) = int32(fd)
 	}
 	return b
 }

vendor/golang.org/x/sys/unix/sockcmsg_unix_other.go

@@ -0,0 +1,38 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build aix darwin freebsd linux netbsd openbsd solaris
+
+package unix
+
+import (
+	"runtime"
+)
+
+// Round the length of a raw sockaddr up to align it properly.
+func cmsgAlignOf(salen int) int {
+	salign := SizeofPtr
+
+	// dragonfly needs to check ABI version at runtime, see cmsgAlignOf in
+	// sockcmsg_dragonfly.go
+	switch runtime.GOOS {
+	case "aix":
+		// There is no alignment on AIX.
+		salign = 1
+	case "darwin", "illumos", "solaris":
+		// NOTE: It seems like 64-bit Darwin, Illumos and Solaris
+		// kernels still require 32-bit aligned access to network
+		// subsystem.
+		if SizeofPtr == 8 {
+			salign = 4
+		}
+	case "netbsd", "openbsd":
+		// NetBSD and OpenBSD armv7 require 64-bit alignment.
+		if runtime.GOARCH == "arm" {
+			salign = 8
+		}
+	}
+
+	return (salen + salign - 1) & ^(salign - 1)
+}

vendor/golang.org/x/sys/unix/syscall_aix.go

@@ -350,49 +350,12 @@ func (w WaitStatus) Signal() Signal {
 
 func (w WaitStatus) Continued() bool { return w&0x01000000 != 0 }
 
-func (w WaitStatus) CoreDump() bool { return w&0x200 != 0 }
+func (w WaitStatus) CoreDump() bool { return w&0x80 == 0x80 }
 
 func (w WaitStatus) TrapCause() int { return -1 }
 
 //sys	ioctl(fd int, req uint, arg uintptr) (err error)
 
-// ioctl itself should not be exposed directly, but additional get/set
-// functions for specific types are permissible.
-
-// IoctlSetInt performs an ioctl operation which sets an integer value
-// on fd, using the specified request number.
-func IoctlSetInt(fd int, req uint, value int) error {
-	return ioctl(fd, req, uintptr(value))
-}
-
-func ioctlSetWinsize(fd int, req uint, value *Winsize) error {
-	return ioctl(fd, req, uintptr(unsafe.Pointer(value)))
-}
-
-func ioctlSetTermios(fd int, req uint, value *Termios) error {
-	return ioctl(fd, req, uintptr(unsafe.Pointer(value)))
-}
-
-// IoctlGetInt performs an ioctl operation which gets an integer value
-// from fd, using the specified request number.
-func IoctlGetInt(fd int, req uint) (int, error) {
-	var value int
-	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
-	return value, err
-}
-
-func IoctlGetWinsize(fd int, req uint) (*Winsize, error) {
-	var value Winsize
-	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
-	return &value, err
-}
-
-func IoctlGetTermios(fd int, req uint) (*Termios, error) {
-	var value Termios
-	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
-	return &value, err
-}
-
 // fcntl must never be called with cmd=F_DUP2FD because it doesn't work on AIX
 // There is no way to create a custom fcntl and to keep //sys fcntl easily,
 // Therefore, the programmer must call dup2 instead of fcntl in this case.

vendor/golang.org/x/sys/unix/syscall_aix_ppc.go

@@ -29,6 +29,10 @@ func (msghdr *Msghdr) SetControllen(length int) {
 	msghdr.Controllen = uint32(length)
 }
 
+func (msghdr *Msghdr) SetIovlen(length int) {
+	msghdr.Iovlen = int32(length)
+}
+
 func (cmsg *Cmsghdr) SetLen(length int) {
 	cmsg.Len = uint32(length)
 }

vendor/golang.org/x/sys/unix/syscall_aix_ppc64.go

@@ -29,6 +29,10 @@ func (msghdr *Msghdr) SetControllen(length int) {
 	msghdr.Controllen = uint32(length)
 }
 
+func (msghdr *Msghdr) SetIovlen(length int) {
+	msghdr.Iovlen = int32(length)
+}
+
 func (cmsg *Cmsghdr) SetLen(length int) {
 	cmsg.Len = uint32(length)
 }

vendor/golang.org/x/sys/unix/syscall_bsd.go

@@ -237,7 +237,7 @@ func anyToSockaddr(fd int, rsa *RawSockaddrAny) (Sockaddr, error) {
 				break
 			}
 		}
-		bytes := (*[10000]byte)(unsafe.Pointer(&pp.Path[0]))[0:n]
+		bytes := (*[len(pp.Path)]byte)(unsafe.Pointer(&pp.Path[0]))[0:n]
 		sa.Name = string(bytes)
 		return sa, nil
 
@@ -413,8 +413,6 @@ func Kevent(kq int, changes, events []Kevent_t, timeout *Timespec) (n int, err e
 	return kevent(kq, change, len(changes), event, len(events), timeout)
 }
 
-//sys	sysctl(mib []_C_int, old *byte, oldlen *uintptr, new *byte, newlen uintptr) (err error) = SYS___SYSCTL
-
 // sysctlmib translates name to mib number and appends any additional args.
 func sysctlmib(name string, args ...int) ([]_C_int, error) {
 	// Translate name to mib number.

vendor/golang.org/x/sys/unix/syscall_darwin.1_12.go

@@ -0,0 +1,29 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build darwin,go1.12,!go1.13
+
+package unix
+
+import (
+	"unsafe"
+)
+
+func Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) {
+	// To implement this using libSystem we'd need syscall_syscallPtr for
+	// fdopendir. However, syscallPtr was only added in Go 1.13, so we fall
+	// back to raw syscalls for this func on Go 1.12.
+	var p unsafe.Pointer
+	if len(buf) > 0 {
+		p = unsafe.Pointer(&buf[0])
+	} else {
+		p = unsafe.Pointer(&_zero)
+	}
+	r0, _, e1 := Syscall6(SYS_GETDIRENTRIES64, uintptr(fd), uintptr(p), uintptr(len(buf)), uintptr(unsafe.Pointer(basep)), 0, 0)
+	n = int(r0)
+	if e1 != 0 {
+		return n, errnoErr(e1)
+	}
+	return n, nil
+}

vendor/golang.org/x/sys/unix/syscall_darwin.1_13.go

@@ -0,0 +1,103 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build darwin,go1.13
+
+package unix
+
+import "unsafe"
+
+//sys	closedir(dir uintptr) (err error)
+//sys	readdir_r(dir uintptr, entry *Dirent, result **Dirent) (res Errno)
+
+func fdopendir(fd int) (dir uintptr, err error) {
+	r0, _, e1 := syscall_syscallPtr(funcPC(libc_fdopendir_trampoline), uintptr(fd), 0, 0)
+	dir = uintptr(r0)
+	if e1 != 0 {
+		err = errnoErr(e1)
+	}
+	return
+}
+
+func libc_fdopendir_trampoline()
+
+//go:linkname libc_fdopendir libc_fdopendir
+//go:cgo_import_dynamic libc_fdopendir fdopendir "/usr/lib/libSystem.B.dylib"
+
+func Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) {
+	// Simulate Getdirentries using fdopendir/readdir_r/closedir.
+	const ptrSize = unsafe.Sizeof(uintptr(0))
+
+	// We store the number of entries to skip in the seek
+	// offset of fd. See issue #31368.
+	// It's not the full required semantics, but should handle the case
+	// of calling Getdirentries or ReadDirent repeatedly.
+	// It won't handle assigning the results of lseek to *basep, or handle
+	// the directory being edited underfoot.
+	skip, err := Seek(fd, 0, 1 /* SEEK_CUR */)
+	if err != nil {
+		return 0, err
+	}
+
+	// We need to duplicate the incoming file descriptor
+	// because the caller expects to retain control of it, but
+	// fdopendir expects to take control of its argument.
+	// Just Dup'ing the file descriptor is not enough, as the
+	// result shares underlying state. Use Openat to make a really
+	// new file descriptor referring to the same directory.
+	fd2, err := Openat(fd, ".", O_RDONLY, 0)
+	if err != nil {
+		return 0, err
+	}
+	d, err := fdopendir(fd2)
+	if err != nil {
+		Close(fd2)
+		return 0, err
+	}
+	defer closedir(d)
+
+	var cnt int64
+	for {
+		var entry Dirent
+		var entryp *Dirent
+		e := readdir_r(d, &entry, &entryp)
+		if e != 0 {
+			return n, errnoErr(e)
+		}
+		if entryp == nil {
+			break
+		}
+		if skip > 0 {
+			skip--
+			cnt++
+			continue
+		}
+		reclen := int(entry.Reclen)
+		if reclen > len(buf) {
+			// Not enough room. Return for now.
+			// The counter will let us know where we should start up again.
+			// Note: this strategy for suspending in the middle and
+			// restarting is O(n^2) in the length of the directory. Oh well.
+			break
+		}
+		// Copy entry into return buffer.
+		s := struct {
+			ptr unsafe.Pointer
+			siz int
+			cap int
+		}{ptr: unsafe.Pointer(&entry), siz: reclen, cap: reclen}
+		copy(buf, *(*[]byte)(unsafe.Pointer(&s)))
+		buf = buf[reclen:]
+		n += reclen
+		cnt++
+	}
+	// Set the seek offset of the input fd to record
+	// how many files we've already returned.
+	_, err = Seek(fd, cnt, 0 /* SEEK_SET */)
+	if err != nil {
+		return n, err
+	}
+
+	return n, nil
+}

vendor/golang.org/x/sys/unix/syscall_darwin.go

@@ -339,43 +339,6 @@ func Kill(pid int, signum syscall.Signal) (err error) { return kill(pid, int(sig
 
 //sys	ioctl(fd int, req uint, arg uintptr) (err error)
 
-// ioctl itself should not be exposed directly, but additional get/set
-// functions for specific types are permissible.
-
-// IoctlSetInt performs an ioctl operation which sets an integer value
-// on fd, using the specified request number.
-func IoctlSetInt(fd int, req uint, value int) error {
-	return ioctl(fd, req, uintptr(value))
-}
-
-func ioctlSetWinsize(fd int, req uint, value *Winsize) error {
-	return ioctl(fd, req, uintptr(unsafe.Pointer(value)))
-}
-
-func ioctlSetTermios(fd int, req uint, value *Termios) error {
-	return ioctl(fd, req, uintptr(unsafe.Pointer(value)))
-}
-
-// IoctlGetInt performs an ioctl operation which gets an integer value
-// from fd, using the specified request number.
-func IoctlGetInt(fd int, req uint) (int, error) {
-	var value int
-	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
-	return value, err
-}
-
-func IoctlGetWinsize(fd int, req uint) (*Winsize, error) {
-	var value Winsize
-	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
-	return &value, err
-}
-
-func IoctlGetTermios(fd int, req uint) (*Termios, error) {
-	var value Termios
-	err := ioctl(fd, req, uintptr(unsafe.Pointer(&value)))
-	return &value, err
-}
-
 func Uname(uname *Utsname) error {
 	mib := []_C_int{CTL_KERN, KERN_OSTYPE}
 	n := unsafe.Sizeof(uname.Sysname)
@@ -497,7 +460,7 @@ func Sendfile(outfd int, infd int, offset *int64, count int) (written int, err e
 //sys	Revoke(path string) (err error)
 //sys	Rmdir(path string) (err error)
 //sys	Seek(fd int, offset int64, whence int) (newoffset int64, err error) = SYS_LSEEK
-//sys	Select(n int, r *FdSet, w *FdSet, e *FdSet, timeout *Timeval) (err error)
+//sys	Select(nfd int, r *FdSet, w *FdSet, e *FdSet, timeout *Timeval) (n int, err error)
 //sys	Setegid(egid int) (err error)
 //sysnb	Seteuid(euid int) (err error)
 //sysnb	Setgid(gid int) (err error)

vendor/golang.org/x/sys/unix/syscall_darwin_386.1_11.go

@@ -0,0 +1,9 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build darwin,386,!go1.12
+
+package unix
+
+//sys  Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) = SYS_GETDIRENTRIES64

vendor/golang.org/x/sys/unix/syscall_darwin_386.go

@@ -10,6 +10,7 @@ import (
 	"syscall"
 )
 
+//sys	sysctl(mib []_C_int, old *byte, oldlen *uintptr, new *byte, newlen uintptr) (err error) = SYS___SYSCTL
 //sys   ptrace(request int, pid int, addr uintptr, data uintptr) (err error)
 
 func setTimespec(sec, nsec int64) Timespec {
@@ -45,6 +46,10 @@ func (msghdr *Msghdr) SetControllen(length int) {
 	msghdr.Controllen = uint32(length)
 }
 
+func (msghdr *Msghdr) SetIovlen(length int) {
+	msghdr.Iovlen = int32(length)
+}
+
 func (cmsg *Cmsghdr) SetLen(length int) {
 	cmsg.Len = uint32(length)
 }
@@ -58,7 +63,6 @@ const SYS___SYSCTL = SYS_SYSCTL
 //sys	Fstat(fd int, stat *Stat_t) (err error) = SYS_FSTAT64
 //sys	Fstatat(fd int, path string, stat *Stat_t, flags int) (err error) = SYS_FSTATAT64
 //sys	Fstatfs(fd int, stat *Statfs_t) (err error) = SYS_FSTATFS64
-//sys	Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) = SYS_GETDIRENTRIES64
 //sys	getfsstat(buf unsafe.Pointer, size uintptr, flags int) (n int, err error) = SYS_GETFSSTAT64
 //sys	Lstat(path string, stat *Stat_t) (err error) = SYS_LSTAT64
 //sys	Stat(path string, stat *Stat_t) (err error) = SYS_STAT64

vendor/golang.org/x/sys/unix/syscall_darwin_amd64.1_11.go

@@ -0,0 +1,9 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build darwin,amd64,!go1.12
+
+package unix
+
+//sys  Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) = SYS_GETDIRENTRIES64

vendor/golang.org/x/sys/unix/syscall_darwin_amd64.go

@@ -10,6 +10,7 @@ import (
 	"syscall"
 )
 
+//sys	sysctl(mib []_C_int, old *byte, oldlen *uintptr, new *byte, newlen uintptr) (err error) = SYS___SYSCTL
 //sys   ptrace(request int, pid int, addr uintptr, data uintptr) (err error)
 
 func setTimespec(sec, nsec int64) Timespec {
@@ -45,6 +46,10 @@ func (msghdr *Msghdr) SetControllen(length int) {
 	msghdr.Controllen = uint32(length)
 }
 
+func (msghdr *Msghdr) SetIovlen(length int) {
+	msghdr.Iovlen = int32(length)
+}
+
 func (cmsg *Cmsghdr) SetLen(length int) {
 	cmsg.Len = uint32(length)
 }
@@ -58,7 +63,6 @@ const SYS___SYSCTL = SYS_SYSCTL
 //sys	Fstat(fd int, stat *Stat_t) (err error) = SYS_FSTAT64
 //sys	Fstatat(fd int, path string, stat *Stat_t, flags int) (err error) = SYS_FSTATAT64
 //sys	Fstatfs(fd int, stat *Statfs_t) (err error) = SYS_FSTATFS64
-//sys	Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) = SYS_GETDIRENTRIES64
 //sys	getfsstat(buf unsafe.Pointer, size uintptr, flags int) (n int, err error) = SYS_GETFSSTAT64
 //sys	Lstat(path string, stat *Stat_t) (err error) = SYS_LSTAT64
 //sys	Stat(path string, stat *Stat_t) (err error) = SYS_STAT64

vendor/golang.org/x/sys/unix/syscall_darwin_arm.1_11.go

@@ -0,0 +1,11 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build darwin,386,!go1.12
+
+package unix
+
+func Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) {
+	return 0, ENOSYS
+}

vendor/golang.org/x/sys/unix/syscall_darwin_arm.go

@@ -12,6 +12,10 @@ func ptrace(request int, pid int, addr uintptr, data uintptr) error {
 	return ENOTSUP
 }
 
+func sysctl(mib []_C_int, old *byte, oldlen *uintptr, new *byte, newlen uintptr) error {
+	return ENOTSUP
+}
+
 func setTimespec(sec, nsec int64) Timespec {
 	return Timespec{Sec: int32(sec), Nsec: int32(nsec)}
 }
@@ -45,6 +49,10 @@ func (msghdr *Msghdr) SetControllen(length int) {
 	msghdr.Controllen = uint32(length)
 }
 
+func (msghdr *Msghdr) SetIovlen(length int) {
+	msghdr.Iovlen = int32(length)
+}
+
 func (cmsg *Cmsghdr) SetLen(length int) {
 	cmsg.Len = uint32(length)
 }
@@ -62,7 +70,3 @@ const SYS___SYSCTL = SYS_SYSCTL
 //sys	Lstat(path string, stat *Stat_t) (err error)
 //sys	Stat(path string, stat *Stat_t) (err error)
 //sys	Statfs(path string, stat *Statfs_t) (err error)
-
-func Getdirentries(fd int, buf []byte, basep *uintptr) (n int, err error) {
-	return 0, ENOSYS
-}