On Sat, Jan 17, 2026 at 12:00:40PM -0800, Otto Kekäläinen wrote:
After uploading a new version of XZ Utils I noticed the autopkgtest
test 'unit-tests-server
' in Debusine is failing. Rerunning the test didn't make it go away,
so it does not seem sporadic.
From
https://ci.debian.net/data/autopkgtest/testing/riscv64/d/debusine/67863465/log.gz
2326s ../../../../usr/lib/python3/dist-packages/_pytest/cacheprovider.py:475
2326s /usr/lib/python3/dist-packages/_pytest/cacheprovider.py:475:
PytestCacheWarning: could not create cache path
/usr/lib/python3/dist-packages/.pytest_cache/v/cache/nodeids: [Errno
13] Permission denied:
'/usr/lib/python3/dist-packages/pytest-cache-files-rtvhjtsp'
2326s config.cache.set("cache/nodeids", sorted(self.cached_nodeids))
2326s
2326s ../../../../usr/lib/python3/dist-packages/_pytest/cacheprovider.py:429
2326s /usr/lib/python3/dist-packages/_pytest/cacheprovider.py:429:
PytestCacheWarning: could not create cache path
/usr/lib/python3/dist-packages/.pytest_cache/v/cache/lastfailed:
[Errno 13] Permission denied:
'/usr/lib/python3/dist-packages/pytest-cache-files-1s1h7mjv'
2326s config.cache.set("cache/lastfailed", self.lastfailed)
The log is very long, not sure if this is the real failure.
It's not. The actual errors start with:
2319s E django.db.utils.OperationalError: out of shared memory
2319s E HINT: You might need to increase "max_locks_per_transaction".
I don't see any relation to how XZ could have broken the test, seems
more plausible that the root cause of the failure is somewhere else
and it just surfaced now triggered by the new test run.
It's indeed not caused by xz-utils: compare
https://ci.debian.net/packages/d/debusine/testing/riscv64/67818322/, for
instance. But it is sporadic in the sense that only one host machine
seems to be affected, so you'll probably find that retrying works
eventually. Looking through recent history:
debci-02: pass x1
debci-31: pass x1
debci-33: pass x1
debci-40: fail x3
debci-bj-04: pass x2, fail x1 but for a different reason
debci-bj-06: pass x2
debci-bj-08: pass x2
So I think debci-40 is misconfigured in some way that causes PostgreSQL
to have trouble. CCing the debian-ci list for help.
--
Colin Watson (he/him) [[email protected]]