Repository navigation
Conversation
Alena0704
force-pushed
the
fix-flaky-tests-rel-2-stable
branch
from
September 29, 2026 16:58
435867e to
335738d
Compare
Alena0704
force-pushed
the
fix-flaky-tests-rel-2-stable
branch
from
October 8, 2026 09:17
335738d to
6a1a193
Compare
Collaborator
Author
|
This pull request is backport fixes of flacky tests for REL_2_STABLE from #2060 |
Alena0704
force-pushed
the
fix-flaky-tests-rel-2-stable
branch
from
October 8, 2026 09:56
d2f12e0 to
462a395
Compare
Alena0704
force-pushed
the
fix-flaky-tests-rel-2-stable
branch
from
October 8, 2026 18:07
462a395 to
0e06bc0
Compare
Collaborator
Author
|
CI failures revealed that the test fixes need adaptation for REL_2_STABLE. I'm reanalyzing commits and tests on REL_2_STABLE. |
Alena0704
marked this pull request as draft
October 8, 2026 18:12
Alena0704
force-pushed
the
fix-flaky-tests-rel-2-stable
branch
3 times, most recently
from
October 8, 2026 22:04
55722c0 to
7571857
Compare
REL_2_STABLE has no SQL function to force pending table statistics to flush. Close each writer session to flush on backend exit, then poll the collector counters with a fresh statistics snapshot on every retry. Restore on_change mode when reconnecting for the third scenario. Wait for analyze, autoanalyze and modification counts before checking the results, and raise an error if the expected state is not reached. Apply the helper and expected results to regular, PAX and single-node tests. Assisted-by: Claude Code Assisted-by: ChatGPT Backpatch-through: REL_2_STABLE
Assisted-by: Claude Code Backpatch-through: REL_2_STABLE
The test already averages CPU samples, but the expected values are still too low. The suite sets the cluster CPU limit to 100%, so expect 100% for one uncapped group and 33%/67% for groups with weights 100/200. Keep the existing tolerance of 10 percentage points. Assisted-by: Claude Code Backpatch-through: REL_2_STABLE
These tasks only check that valid schedules are accepted. Deactivate them after creation to avoid background runs interfering with cleanup and the oid_wraparound test. Assisted-by: Claude Code Backpatch-through: REL_2_STABLE
Close both aborted-insert sessions before polling statistics so pending reports are flushed on backend exit. Wait on segment 1, which contains all rows: pg_stat_all_tables aggregates its tuple counters from segments, so waiting on the coordinator does not synchronize the following read. Keep the statistics check before DELETE to avoid racing its counters. Wait for the second VACUUM count instead of a dead-tuple count that was already zero after the first VACUUM. (cherry picked from commit fba73b0) Assisted-by: ChatGPT
hot_standby/transaction_isolation deliberately panics the standby and reconnects to check distributed transaction visibility after recovery. The new connection can arrive before consistent recovery is reached, when the server reports "the database system is not yet accepting connections". This message was missing from the connection retry list. Retry this transient state using the existing attempt limit and interval. Keep the crash scenario and all transaction visibility checks intact. Backport required: REL_2_STABLE is affected by the same race. (cherry picked from commit 9ec9f4c)
The final termination scenario can receive a disconnect before backend exit has removed its temporary table. Keep pg_terminate_backend() asynchronous and use wait_for_temp_table_cleanup() to poll pg_class until no temporary relations belong to the test role, before DROP ROLE. This prevents a leftover role and resource group from affecting later tests. Bound the retries and report an error if cleanup never completes.
Alena0704
force-pushed
the
fix-flaky-tests-rel-2-stable
branch
from
October 9, 2026 12:33
7571857 to
8f80d90
Compare
tuhaihe
approved these changes
Oct 9, 2026
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.
Collect all fixes from #2032 that can improve CI stability after having reviewed with refactored commit messages.
Author: Dianjin Wang wangdianjin@gmail.com
Assisted-by: Claude Code
Type of Change
Breaking Changes
Test Plan
make installcheckmake -C src/test installcheck-cbdb-parallelImpact
Performance:
User-facing changes:
Dependencies:
Checklist
Additional Context
CI Skip Instructions