Block plan application until state store has caught up to raft #5411
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
schmichael
force-pushed
the
b-snapshotafter
branch
from
983a68c to
0c45898
Compare
dadgar
requested changes
May 15, 2019
preetapan
reviewed
May 17, 2019
Generalize wait for index logic in the state store for reuse elsewhere. Also begin plumbing in a context to combine handling of timeouts and shutdown.
I don't think it's been used for a long time.
Wait for state store to catch up with raft when applying plans.
Avoid returning context.DeadlineExceeded as it lacks helpful information and is often ignored or handled specially by callers.
schmichael
force-pushed
the
b-snapshotafter
branch
from
0c45898 to
a2e4f12
Compare
dadgar
approved these changes
May 20, 2019
schmichael
added a commit
that referenced
this pull request
Jun 3, 2019
Revert plan_apply.go changes from #5411 Since non-Command Raft messages do not update the StateStore index, SnapshotAfter may unnecessarily block and needlessly fail in idle clusters where the last Raft message is a non-Command message. This is trivially reproducible with the dev agent and a job that has 2 tasks, 1 of which fails. The correct logic would be to SnapshotAfter the previous plan's index to ensure consistency. New clusters or newly elected leaders will not have a previous plan, so the index the leader was elected should be used instead.
|
I'm going to lock this pull request because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active contributions. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Do not merge until 0.9.1
This guards against a situation where planApply could dequeue a plan that references objects that are committed to raft but not yet applied to the leader's state store.
To do this I refactored
worker.waitForIndexintoStateStore.SnapshotAfter(new name welcome) so it could be shared between the 3 call sites. Backoff logic should be maintained although error log lines will vary slightly from the old waitForIndex errors ("timeout" vs "deadline exceeded").TODOs