1
0
mirror of https://github.com/kubernetes-sigs/descheduler.git synced 2026-01-26 05:14:13 +01:00
Commit Graph

274 Commits

Author SHA1 Message Date
Shubham Minglani
40bb490f4c add RemovePodsViolatingNodeAffinity strategy
This commit adds requiredDuringSchedulingIgnoredDuringExecution
for RemovePodsViolatingNodeAffinity strategy.

Also adds unit tests and documentation.
2018-02-19 16:20:59 +05:30
Avesh Agarwal
08729f6ef9 Merge pull request #79 from jelmersnoeck/multi-stage-builds
Scratch image + multistage builds.
2018-02-16 12:40:25 -05:00
Jelmer Snoeck
3dd7de8132 Use docker target in README. 2018-02-16 18:30:49 +01:00
Jelmer Snoeck
471aeb5ea4 Set up multistage Docker build.
By setting up a multistage Docker build, we can create the container
image in a single command. This eliminates external setup and allows us
to build this automatically on registries.
2018-02-16 18:30:48 +01:00
Jelmer Snoeck
65e7093ee7 Use scratch image for container.
By disabling CGO, we can use the `scratch` image instead of the `fedora`
image, allowing a lighter weight image.
2018-02-16 08:50:22 +01:00
RaviSantosh Gudimetla
fc0cd4ba30 Merge pull request #78 from ravisantoshgudimetla/schedulable-nodes
Descheduler should not run when cluster size is 1
2018-02-13 17:40:18 -05:00
ravisantoshgudimetla
ba3eac6c57 Descheduler should not run when cluster size is 1 2018-02-08 21:15:02 -05:00
Avesh Agarwal
11a95ce8fb Merge pull request #71 from ravisantoshgudimetla/release-table
Compatibility matrix for kube version
2018-01-29 08:52:59 -05:00
ravisantoshgudimetla
29a9fc6b56 Compatibility matrix for kube versions 2018-01-13 22:25:40 +05:30
Avesh Agarwal
d3c2f25685 Merge pull request #73 from aveshagarwal/master-rebase-kube-1-9
Rebase to kube 1 9
v0.4.0
2018-01-09 13:00:42 -05:00
Avesh Agarwal
e858c9ee80 Fix glog issue: ERROR: logging before flag.Parse. 2018-01-09 12:54:52 -05:00
Avesh Agarwal
44752e5e83 Update code and hack dir for kube 1.9 rebase. 2018-01-09 12:41:01 -05:00
Avesh Agarwal
7123f30783 Update auto-generated code. 2018-01-09 11:04:26 -05:00
Avesh Agarwal
a82cf7cea4 Update verify-gofmt for go1.9. 2018-01-09 11:04:26 -05:00
Avesh Agarwal
72318868b0 Update vendor dir for kube 1.9 rebase. 2018-01-09 00:58:09 -05:00
Avesh Agarwal
589fb95236 Update glide files for kube 1.9 rebase. 2018-01-09 00:57:56 -05:00
Avesh Agarwal
7df543d137 Merge pull request #70 from ravisantoshgudimetla/pod-antiaffinity-readme-update
Read me updated with changes for pod anti-affinity
2018-01-08 15:04:59 -05:00
Avesh Agarwal
1d7f429ba1 Merge pull request #72 from ravisantoshgudimetla/owner-ref-switch
Owner ref switch
2018-01-08 15:03:10 -05:00
ravisantoshgudimetla
bf29a6073f Owner ref switch
Signed-off-by: ravisantoshgudimetla <ravisantoshgudimetla@gmail.com>
2018-01-05 00:22:12 +05:30
ravisantoshgudimetla
65635bdb2e Read me updated with changes for pod anti-affinity 2018-01-03 12:08:15 +05:30
Avesh Agarwal
fd961557d0 Merge pull request #61 from ravisantoshgudimetla/latency-sensitive-pods
Make GPU requested pods not evictable by descheduler
2018-01-02 16:43:27 -05:00
Avesh Agarwal
c29bdc1dbe Merge pull request #64 from ravisantoshgudimetla/fix-serviceaccount
Fixes the serviceaccount issue by adding new eviction resources
2018-01-02 16:40:36 -05:00
RaviSantosh Gudimetla
ffecc54bf5 Merge pull request #68 from ravisantoshgudimetla/go_report
Go report included
2017-12-31 19:40:09 +05:30
ravisantoshgudimetla
f6f6fbab10 Go report included 2017-12-31 19:34:06 +05:30
RaviSantosh Gudimetla
12c217477c Merge pull request #67 from ravisantoshgudimetla/build-badge
Build status badge
2017-12-29 19:11:47 +05:30
ravisantoshgudimetla
20a4798465 Build status badge 2017-12-29 19:09:17 +05:30
RaviSantosh Gudimetla
a201f222e5 Merge pull request #66 from spiffxp/update-code-of-conduct
Update code-of-conduct.md /cc @aveshagarwal
2017-12-26 11:54:32 +05:30
Aaron Crickenberger
6e705fde85 Update code-of-conduct.md
Refer to kubernetes/community as authoritative source for code of conduct
2017-12-20 14:11:52 -05:00
ravisantoshgudimetla
5763554be4 Fixes the serviceaccount issue by adding new eviction resources 2017-12-14 15:13:33 +05:30
ravisantoshgudimetla
6f873d5e69 Make GPU requested pods not evictable by descheduler 2017-12-05 03:49:47 -05:00
Avesh Agarwal
445ae92caa Merge pull request #59 from aveshagarwal/master-fix-errors
Update logging and a few fixes.
2017-12-01 13:19:36 -05:00
Avesh Agarwal
8f3c0cf4b8 Merge pull request #58 from aveshagarwal/master-fix-node-lister
Fix node lister by allowing its reflector to have enough time so that listing works.
2017-12-01 13:02:17 -05:00
Avesh Agarwal
2a280f9a20 Update logging and a few fixes. 2017-12-01 13:02:05 -05:00
Avesh Agarwal
0503f53904 Merge pull request #57 from aveshagarwal/master-fix-pod-listing
Fix field selector to only list non-terminated pods.
2017-12-01 12:58:50 -05:00
Avesh Agarwal
e0a9dfcb76 Fix field selector to only list non-terminated pods. 2017-12-01 12:50:39 -05:00
Avesh Agarwal
5db49f2ce1 Fix node lister by allowing its reflector to have enough time so that listing works.
Currently time duration is chosen randomly.
2017-11-30 12:28:33 -05:00
Avesh Agarwal
582bd67681 Merge pull request #49 from ravisantoshgudimetla/podlist-refactor
Small refactor to accomodate descheduling only evictable pods
2017-11-29 15:48:04 -05:00
ravisantoshgudimetla
a63f815116 Small refactor to accomodate descheduling only evictable pods
Signed-off-by: ravisantoshgudimetla <ravisantoshgudimetla@gmail.com>
2017-11-29 15:34:28 -05:00
Avesh Agarwal
afc17a62ea Merge pull request #45 from ravisantoshgudimetla/lownodeutilization-fix#43
Changes to fix low node utilization strategy
2017-11-27 18:47:42 -05:00
ravisantoshgudimetla
6dbc8a1fcc Changes to fix low node utilization strategy 2017-11-27 18:40:53 -05:00
RaviSantosh Gudimetla
d2bd16a12d Merge pull request #41 from containscafeine/error-out-on-unsupported-resource-names
Validate resource names to pass to thresholds
2017-11-27 12:35:30 -05:00
Avesh Agarwal
955d0eb228 Merge pull request #36 from ravisantoshgudimetla/flag-fix
kubeconfig file correction
2017-11-27 11:15:27 -05:00
Shubham Minglani
a490726245 Validate resource names to pass to thresholds
This commit adds checks to only allow valid resource names for the
thresholds field, which are "cpu", "memory", and "pods" right now.

For the other valid or invalid resource names, descheduler will now
throw an error.

Also, tests have been added to test the added behavior.

fix #40
2017-11-27 19:54:53 +05:30
RaviSantosh Gudimetla
16a504fb87 Merge pull request #44 from containscafeine/log-underutilized-and-target-nodes
Add log messages for target, underutilized nodes
2017-11-26 11:55:26 -05:00
Shubham Minglani
d54b73a6ba Add log messages for target, underutilized nodes
This commit adds log messages to be printed whenever a node is
identified as a target node from which the pods can be evicted or
as an underutilized node to which the evicted pods can be
scheduled, if at all.

This should help with debugging as well as with letting the user
know the identified nodes.
2017-11-25 19:49:05 +05:30
ravisantoshgudimetla
344dc0f3c2 kubeconfig file correction 2017-11-20 22:40:15 -05:00
Avesh Agarwal
9f38146bbf Merge pull request #34 from aveshagarwal/master-remove-deadcode
Remove unused test code.
v0.3.0
2017-11-10 17:06:05 -05:00
Avesh Agarwal
1473e1d024 Remove unused test code. 2017-11-10 16:18:07 -05:00
Avesh Agarwal
7b4b9d9e7e Merge pull request #35 from aveshagarwal/master-fixes
Fix to not process empty node list and also fix error reporting in other places.
2017-11-10 16:14:11 -05:00
Avesh Agarwal
7d079813e5 Fix to not process empty node list and also fix error reporting in other places.
Also fix unit test panic and if nodeLister is nil and refactor some code.
2017-11-10 16:09:33 -05:00