Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

roachtest/mixedversion: TestTestPlanner should restore default version #140753

Conversation

srosenberg
Copy link
Member

We saw that Test_maxNumPlanSteps suddenly failed in an otherwise unrelated backport [1]. The reason turned out to be non-determinisim. That is, a different
unit test, namely TestTestPlanner did not
restore the default version (clusterupgrade.TestBuildVersion).

This change forwardports the missing restore to
ensure future test executions follow the same
PRNG sequence.

[1] #140674

Epic: none

Release note: None

We saw that `Test_maxNumPlanSteps` suddenly failed in an
otherwise unrelated backport [1]. The reason turned
out to be non-determinisim. That is, a different
unit test, namely `TestTestPlanner` did not
restore the default version (`clusterupgrade.TestBuildVersion`).

This change forwardports the missing restore to
ensure future test executions follow the same
PRNG sequence.

[1] cockroachdb#140674

Epic: none

Release note: None
@srosenberg srosenberg requested a review from a team as a code owner February 7, 2025 23:42
@srosenberg srosenberg requested review from herkolategan and DarrylWong and removed request for a team February 7, 2025 23:42
@cockroach-teamcity
Copy link
Member

This change is Reviewable

@srosenberg srosenberg added the backport-25.1.x Flags PRs that need to be backported to 25.1 label Feb 7, 2025
@srosenberg
Copy link
Member Author

TFTR!

bors r=herkolategan,darrylwon

craig bot pushed a commit that referenced this pull request Feb 10, 2025
140753: roachtest/mixedversion: TestTestPlanner should restore default version r=herkolategan,darrylwon a=srosenberg

We saw that `Test_maxNumPlanSteps` suddenly failed in an otherwise unrelated backport [1]. The reason turned out to be non-determinisim. That is, a different
unit test, namely `TestTestPlanner` did not
restore the default version (`clusterupgrade.TestBuildVersion`).

This change forwardports the missing restore to
ensure future test executions follow the same
PRNG sequence.

[1] #140674

Epic: none

Release note: None

Co-authored-by: Stan Rosenberg <[email protected]>
@craig
Copy link
Contributor

craig bot commented Feb 10, 2025

Build failed:

@srosenberg
Copy link
Member Author

bors retry

craig bot pushed a commit that referenced this pull request Feb 10, 2025
140753: roachtest/mixedversion: TestTestPlanner should restore default version r=herkolategan,darrylwon a=srosenberg

We saw that `Test_maxNumPlanSteps` suddenly failed in an otherwise unrelated backport [1]. The reason turned out to be non-determinisim. That is, a different
unit test, namely `TestTestPlanner` did not
restore the default version (`clusterupgrade.TestBuildVersion`).

This change forwardports the missing restore to
ensure future test executions follow the same
PRNG sequence.

[1] #140674

Epic: none

Release note: None

140785: rac2,kvserver: start the StreamCloseScheduler r=sumeerbhola a=pav-kv

This commit enables the `StreamCloseScheduler`, which is responsible for closing RACv2 streams some time (400ms) after they enter `StateProbe`.

The initialization had to move from `NewStore` to `Store.Start()` because it needs the `stopper` to start the job.

Epic: none
Release note: none

140791: sem/tree: avoid assertion error on unimplemented builtins in views r=yuzefovich a=yuzefovich

Previously, we would hit an assertion error when trying to use an unimplemented builtin in the CREATE VIEW statement and this is now fixed. The issue is that we resolve unimplemented builtins as a definition with zero overloads, and all places that previously assumed at least one existing overload have been audited. The resolved function definition has been updated to have "unsupported with issue" integer indicating why there are no overloads.

Additionally, `UnsupportedWithIssue` property is now changed to be `uint` since we didn't use "negative value as having no corresponding issue" ability.

I decided to not include a release note since this seems like an edge case and we've only seen this a handful of times in sentry.

Fixes: #128535.

Release note: None

141009: sqlstats: use `BatchProcessLatencyBuckets` for flush latency r=xinhaoz a=dhartunian

The max of 10s on the IO latency buckets is too short to measure flush latency effectively. Previously, this metric was measuring per-statement flush latency, but this was altered in #122919.

Release note: None

Co-authored-by: Stan Rosenberg <[email protected]>
Co-authored-by: Pavel Kalinnikov <[email protected]>
Co-authored-by: Yahor Yuzefovich <[email protected]>
Co-authored-by: David Hartunian <[email protected]>
@craig
Copy link
Contributor

craig bot commented Feb 10, 2025

Build failed (retrying...):

@craig craig bot merged commit c38b185 into cockroachdb:master Feb 10, 2025
22 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
backport-25.1.x Flags PRs that need to be backported to 25.1
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants