sql: fix bug whereby backfiller would drop spans on txn restart #54755
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
This bug was caught by testing with cockroachdb#54695. Before that change, it would fail almost immediately, now it does not fail under stress. I'm open to suggestions on how to more generally test this. Release note (bug fix): Fixed a rare bug which can lead to index backfills failing in the face of transaction restarts.
|
This change is |
thoszhang
approved these changes
Sep 24, 2020
There was a problem hiding this comment.
LGTM but I think someone on Bulk IO should take a look too.
What are the errors you can get from this? Does this ever fail at index validation?
Reviewed 1 of 1 files at r1.
Reviewable status:complete! 0 of 0 LGTMs obtained
thoszhang
requested review from
a team and
miretskiy
and removed request for
a team
|
(@miretskiy I just added cockroachdb/bulk-prs but feel free to reassign to someone else) |
|
The main error I was seeing was: |
|
Thanks. I wonder if the panic specifically is due to an inconsistency in in-memory state that goes away if the job restarts, and if there are more subtle possible corruption bugs that don't manifest an immediate crash in the backfiller. I think if that happens, we'd catch it in the validation step, but it would be ideal to confirm that. |
I'm pretty sure the operation of subtracting spans ends up being idempotent so we shouldn't have persistent correctness problems due to this one. The issue here is just that we might create an empty set of spans and then upon retry hit the panic. Given I don't see any sentry reports on this one, we might just be in luck |
miretskiy
approved these changes
Sep 24, 2020
There was a problem hiding this comment.
Reviewable status:
|
bors r+ |
|
Build succeeded: |
This was referenced Sep 28, 2020
ajwerner
added a commit
to ajwerner/cockroach
that referenced
this pull request
Sep 28, 2020
Fixes a bug introduced in cockroachdb#54755 whereby we'd always subtract from the original set of spans rather than from the updated set of spans meaning that we could backfill the same span multiple times. Fixes cockroachdb#54775. Release note: None
This was referenced Sep 28, 2020
craig bot
pushed a commit
that referenced
this pull request
Sep 29, 2020
54900: sql: fix a bug tracking spans in a backfill r=lucy-zhang a=ajwerner Fixes a bug introduced in #54755 whereby we'd always subtract from the original set of spans rather than from the updated set of spans meaning that we could backfill the same span multiple times. Fixes #54775. Release note: None Co-authored-by: Andrew Werner <[email protected]>
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 bug was caught by testing with #54695. Before that change, it would fail
almost immediately, now it does not fail under stress. I'm open to suggestions
on how to more generally test this.
Release note (bug fix): Fixed a rare bug which can lead to index backfills
failing in the face of transaction restarts.