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

Fix flaky append regression test #7512

Merged
merged 1 commit into from
Dec 3, 2024

Conversation

fabriziomello
Copy link
Contributor

@fabriziomello fabriziomello commented Dec 3, 2024

Fixed it by manually running VACUUM ANALYZE in all involved hypertables to avoid flaky output tests.

https://github.com/timescale/timescaledb/actions/runs/12128976291/job/33819502944?pr=7505#step:17:23

Disable-check: force-changelog-file

@fabriziomello fabriziomello added testing planner flaky-test Issue about a flaky test labels Dec 3, 2024
@fabriziomello fabriziomello self-assigned this Dec 3, 2024
@fabriziomello fabriziomello changed the title Fix flaky append Fix flaky append regression test Dec 3, 2024
@github-actions github-actions bot requested review from akuzm and erimatnor December 3, 2024 00:34
Copy link

github-actions bot commented Dec 3, 2024

@erimatnor, @akuzm: please review this pull request.

Powered by pull-review

Copy link
Contributor

@mkindahl mkindahl left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Although this works fine in most cases, it makes it impossible to, for example, test for the number of heap fetches in particular tests.

Approving since I can't see that it is something we are doing right now, so we can deal with it when the need arises.

@akuzm
Copy link
Member

akuzm commented Dec 3, 2024

Shouldn't this be fixed by vacuum as well? IIRC we had a similar problem for transparent decompression test

@fabriziomello fabriziomello force-pushed the fix_flaky_append branch 2 times, most recently from bd9401a to ca7c4d0 Compare December 3, 2024 15:39
@fabriziomello
Copy link
Contributor Author

fabriziomello commented Dec 3, 2024

Shouldn't this be fixed by vacuum as well? IIRC we had a similar problem for transparent decompression test

I've sent another commit executing a manual VACUUM and it changed a lot the output: ca7c4d0

At least all the Heap Fetches now are zero :-)

Copy link

codecov bot commented Dec 3, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 82.18%. Comparing base (59f50f2) to head (dc4cde3).
Report is 640 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #7512      +/-   ##
==========================================
+ Coverage   80.06%   82.18%   +2.11%     
==========================================
  Files         190      230      +40     
  Lines       37181    43191    +6010     
  Branches     9450    10869    +1419     
==========================================
+ Hits        29770    35495    +5725     
- Misses       2997     3371     +374     
+ Partials     4414     4325      -89     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Fixed it by manually running `VACUUM ANALYZE` in all involved
hypertables to avoid flaky output tests.
@fabriziomello fabriziomello enabled auto-merge (rebase) December 3, 2024 18:50
@fabriziomello fabriziomello merged commit f1d201e into timescale:main Dec 3, 2024
52 checks passed
fabriziomello added a commit to fabriziomello/timescaledb that referenced this pull request Dec 5, 2024
In timescale#7512 we added manual `VACUUM` in all involved relations to avoid
flaky output tests.

But looks like it is not working very well specially on Windows builds,
so now we're replacing the output `Heap Fetches: [0-9]` for `Heap
Fetches: 0` to make have a more predictable output test.
fabriziomello added a commit to fabriziomello/timescaledb that referenced this pull request Dec 5, 2024
In timescale#7512 we added manual `VACUUM` in all involved relations to avoid
flaky output tests.

But looks like it is not working very well specially on Windows builds,
so now we're replacing the output `Heap Fetches: [0-9]` for `Heap
Fetches: 0` to make have a more predictable output test.
fabriziomello added a commit to fabriziomello/timescaledb that referenced this pull request Dec 5, 2024
In timescale#7512 we added manual `VACUUM` in all involved relations to avoid
flaky output tests.

But looks like it is not working very well specially on Windows builds,
so now we're replacing the output `Heap Fetches: [0-9]` for `Heap
Fetches: 0` to make have a more predictable output test.
fabriziomello added a commit to fabriziomello/timescaledb that referenced this pull request Dec 6, 2024
In timescale#7512 we added manual `VACUUM` in all involved relations to avoid
flaky output tests.

But looks like it is not working very well specially on Windows builds,
so now we're replacing the output `Heap Fetches: [0-9]` for `Heap
Fetches: 0` to make have a more predictable output test.
fabriziomello added a commit that referenced this pull request Dec 6, 2024
In #7512 we added manual `VACUUM` in all involved relations to avoid
flaky output tests.

But looks like it is not working very well specially on Windows builds,
so now we're replacing the output `Heap Fetches: [0-9]` for `Heap
Fetches: 0` to make have a more predictable output test.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
flaky-test Issue about a flaky test planner testing
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants