pghoard
pghoard copied to clipboard
tests: increase timeout when waiting for more xlog
About this change - What it does
Make the waiting time for xlog during tests proportional to the number of xlog.
Why this way
test_pause_on_disk_full
was waiting for more xlog than other tests and would timeout while waiting despite having done what the test expected (pause/resume).
Codecov Report
Merging #533 (c5957a9) into main (370b1f3) will decrease coverage by
0.27%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## main #533 +/- ##
==========================================
- Coverage 90.76% 90.49% -0.28%
==========================================
Files 30 30
Lines 4593 4567 -26
==========================================
- Hits 4169 4133 -36
- Misses 424 434 +10
Impacted Files | Coverage Δ | |
---|---|---|
pghoard/walreceiver.py | 93.57% <0.00%> (-3.58%) |
:arrow_down: |
pghoard/webserver.py | 89.04% <0.00%> (-1.03%) |
:arrow_down: |
pghoard/gnutaremu.py | 82.05% <0.00%> (-0.34%) |
:arrow_down: |
pghoard/basebackup/base.py | 85.32% <0.00%> (-0.27%) |
:arrow_down: |
pghoard/compressor.py | 92.95% <0.00%> (-0.13%) |
:arrow_down: |
pghoard/common.py | 93.33% <0.00%> (-0.08%) |
:arrow_down: |
pghoard/archive_sync.py | 86.44% <0.00%> (-0.08%) |
:arrow_down: |
pghoard/transfer.py | 98.64% <0.00%> (-0.03%) |
:arrow_down: |
pghoard/pghoard.py | 82.68% <0.00%> (-0.03%) |
:arrow_down: |
pghoard/basebackup/chunks.py | 96.75% <0.00%> (-0.03%) |
:arrow_down: |
... and 2 more |
The test runs very fast, so actually the timeout increase does not bring much. The problem seems to be in walreceiver, which can't finalize last WAL segment even though it's finished, unless some activity happens after new segment is created. https://github.com/postgres/postgres/commit/596ba75cb11173a528c6b6ec0142a282e42b69ec
I've made a fix for 2.2.2 in https://github.com/aiven/pghoard/pull/552 and will prepare the same for main. I guess this PR can be closed.