Fix flaky append regression test

[fabriziomello]

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

[github-actions]

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

Powered by pull-review

[mkindahl]

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]

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

[fabriziomello]

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 :-)

[codecov]

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.