tests: longer timeouts in test_timeline_deletion_with_files_stuck_in_upload_queue (#8438)

## Problem

This test had two locations with 2 second timeouts, which is rather low
when we run on a highly contended test machine running lots of tests in
parallel. It usually passes, but today I've seen both of these locations
time out on separate PRs.

Example failure:
https://neon-github-public-dev.s3.amazonaws.com/reports/pr-8432/10007868041/index.html#suites/837740b64a53e769572c4ed7b7a7eeeb/6c6a092be083d27c

## Summary of changes

- Change 2 second timeouts to 20 second timeouts
This commit is contained in:
John Spray
2024-07-19 18:30:28 +01:00
committed by Christian Schwarz
parent 067363fe95
commit bba062e262

View File

@@ -577,7 +577,7 @@ def test_timeline_deletion_with_files_stuck_in_upload_queue(
> 0
)
wait_until(20, 0.1, assert_compacted_and_uploads_queued)
wait_until(200, 0.1, assert_compacted_and_uploads_queued)
# Regardless, give checkpoint some time to block for good.
# Not strictly necessary, but might help uncover failure modes in the future.
@@ -619,7 +619,7 @@ def test_timeline_deletion_with_files_stuck_in_upload_queue(
)
# timeline deletion should be unblocking checkpoint ops
checkpoint_thread.join(2.0)
checkpoint_thread.join(20.0)
assert not checkpoint_thread.is_alive()
# Just to be sure, unblock ongoing uploads. If the previous assert was incorrect, or the prometheus metric broken,