Fix flakey TestPluginCleanup in CI #49104
Closed
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
Acquire lock for all Access List resources before locking individual Access List in `ListAccessListMembers`. What seemed to happen in CI was `ListAccessListMembers` was called, thereby locking the individual list. If IneligibilityReconciler ran simultaneously, it would lock all Access Lists, then fail to acquire the individual Access List's lock, leading to a deadlock. This didn't seem to come up in real-world conditions, but was causing flakey failures in CI when running many times at once.
kiosion
added
the
no-changelog
zmb3
reviewed
Nov 17, 2024
| err := a.service.RunWhileLocked(ctx, lockName(accessListName), accessListLockTTL, func(ctx context.Context, _ backend.Backend) error { | ||
| _, err := a.service.GetResource(ctx, accessListName) | ||
| if err != nil { | ||
| err := a.service.RunWhileLocked(ctx, []string{accessListResourceLockName}, accessListLockTTL, func(ctx context.Context, _ backend.Backend) error { |
There was a problem hiding this comment.
So only one auth server can ever list access list members at a time?
Any scale concerns here?
There was a problem hiding this comment.
Possibly... Going over this in more detail today; I'm not really satisfied with this fix. It might work to not lock all lists on update/upsert of a single list instead.
|
If this deadlock can only happen during testing and not in real-world scenarios:
why not align the tests accordingly and leave the implementation unchanged? |
|
Found the actual cause – opened #5530. |
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Outdated suggestions cannot be applied.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.
Acquire lock for all Access List resources before locking individual Access List in
ListAccessListMembers.What seemed to happen in CI is if
ListAccessListMemberswas called, it would lock the individual list. If IneligibilityReconciler then ran simultaneously, it would lock all Access Lists, before failing to acquire the individual Access List's lock, leading to a deadlock.This didn't seem to come up in real-world conditions, but was causing flakey failures in CI when running many times at once.
Resolves #5456.