Prevent runners from being stuck in Terminating when pod disappeared without standard termination process #1318
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…t standard termination process This fixes the said issue by additionally treating any runner pod whose phase is Failed or the runner container exited with non-zero code as "complete" so that ARC gives up unregistering the runner from Actions, deletes the runner pod anyway. Note that there are a plenty of causes for that. If you are deploying runner pods on AWS spot instances or GCE preemptive instances and a job assigned to a runner took more time than the shutdown grace period provided by your cloud provider (2 minutes for AWS spot instances), the runner pod would be terminated prematurely without letting actions/runner unregisters itself from Actions. If your VM or hypervisor failed then runner pods that were running on the node will become PodFailed without unregistering runners from Actions. Please beware that it is currently users responsibility to clean up any dangling runner resources on GitHub Actions. Ref #1307 Might also relate to #1273
mumoshu
changed the title
2 tasks
2 tasks
billimek
added a commit
to homedepot/actions-runner-controller
that referenced
this pull request
Apr 12, 2022
* master: (30 commits) chore(deps): update azure/setup-helm action to v2.1 (actions#1328) [skip ci] Add more usages of RUNNER_VERSION to Renovate config. (actions#1313) chore: fix typo (actions#1316) [skip ci] chore: bump chart to latest (actions#1319) Fix release workflow Prevent runners from stuck in Terminating when pod disappeared without standard termination process (actions#1318) ci: pin go version to the known working version (actions#1303) chore: bump chart to latest (actions#1300) Fix runner pod to be cleaned up earlier regardless of the sync period (actions#1299) Make the hard-coded runner startup timeout to avoid race on token expiration longer (actions#1296) docs: highlight why persistent are not ideal fix(deps): update module sigs.k8s.io/controller-runtime to v0.11.2 chore(deps): update dependency actions/runner to v2.289.2 chore(deps): update helm/chart-releaser-action action to v1.4.0 (actions#1287) refactor(runner/entrypoint): check for externalstmp (actions#1277) docs: redundant words docs: wording docs: add autoscaling also causes problems chore: new line for consistency docs: use the right font ...
2 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This fixes the said issue by additionally treating any runner pod whose phase is Failed or the runner container exited with non-zero code as "complete" so that ARC gives up unregistering the runner from Actions, deletes the runner pod anyway.
Note that there might be plenty of possible causes for that. If you are deploying runner pods on AWS spot instances or GCE preemptive instances and a job assigned to a runner took more time than the shutdown grace period provided by your cloud provider (2 minutes for AWS spot instances), the runner pod would be terminated prematurely without letting actions/runner unregisters itself from Actions. If your VM or hypervisor failed then runner pods that were running on the node will become PodFailed without unregistering runners from Actions.
Please beware that ARC leaves dangling runners on Actions in such circumstances, with or without this patch, because ARC can't unregister runners because Actions doesn't automatically change disappeared runners from busy to non-busy, at least for several hours, and there's no API to forcefully change runners to non-busy or force-delete runners(DeleteRunner API returns 422 on busy runners). It is currently the users' responsibility to clean up any dangling runner resources on GitHub Actions.
Ref #1307
Might also relate to #1273