sql: disable VC injection in TestTemporaryObjectCleaner#169914
Merged
trunk-io[bot] merged 1 commit intocockroachdb:masterfrom May 8, 2026
Merged
Conversation
The test wires `TempObjectsCleanupCh` and `OnTempObjectsCleanupDone` testing knobs into the cleaner, sends one trigger per node per cycle, and expects to read one completion per node. With an auto-injected virtual cluster, both the system tenant and the application layer run a `TemporaryObjectCleaner` per pod and they share the same testing knobs, so 6 cleaners listen on a channel that only delivers 3 triggers per cycle. Worse, the application-layer cleaners run with no coordination (only the system tenant is gated on the meta1 leaseholder), and concurrent application-layer cleaners can race on `ListSessions` when a pod's SQL instance reader cache is briefly stale, deleting the still-active session's temp schema. The race in the application-layer cleanup path is the real bug exposed by this test; it's tracked separately in cockroachdb#169912. The test itself was written for the system-tenant single-cleaner-per-cycle path, so disable VC injection here and let cockroachdb#169912 drive the production fix. Fixes cockroachdb#169663. Epic: none Release note: None
Contributor
|
😎 Merged successfully - details. |
Member
|
This change is |
mw5h
approved these changes
May 8, 2026
Contributor
mw5h
left a comment
mw5h
left a comment
There was a problem hiding this comment.
@mw5h reviewed all commit messages and made 1 comment.
Reviewable status:complete! 1 of 0 LGTMs obtained.
Collaborator
Author
|
TFTR /trunk merge |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The test wires a shared cleanup channel through the
TempObjectsCleanupChtesting knob and sends one trigger per node per cycle, expecting one
completion per node. With an auto-injected virtual cluster, both the
system tenant and the application layer run a
TemporaryObjectCleanerper pod and they share the same testing knobs — 6 cleaners listen on
a channel that only delivers 3 triggers per cycle. Worse, the
application-layer cleaners run with no coordination (only the system
tenant is gated on the meta1 leaseholder), and concurrent
application-layer cleaners can race on
ListSessionswhen a pod'sSQL instance reader cache is briefly stale, deleting the still-active
session's temp schema.
The race in the application-layer cleanup path is the real bug
exposed by this test; it's tracked separately in #169912 (production
fix work). The test itself was written for the system-tenant
single-cleaner-per-cycle path, so disable VC injection here and let
#169912 drive the production fix.
Fixes #169663.
Epic: none
Release note: None