Re: Failures in constraints regression test, "read only 0 of 8192 bytes"

On 3/8/24 09:33, Thomas Munro wrote:
> Happened again. I see this is OpenSUSE. Does that mean the file
> system is Btrfs?

It is, but I don't think that matters - I've been able to reproduce this
locally on my laptop using ext4 filesystem. I'd bet the important piece
here is -DCLOBBER_CACHE_ALWAYS (and it seems avocet/trilobite are the
only animals running with this).

Also, if this really is a filesystem (or environment) issue, it seems
very strange it'd only affect HEAD and not the other branches. So it
seems quite likely this is actually triggered by a commit.

Looking at the commits from the good/bad runs, I see this:

avocet: good=4c2369a bad=f5a465f
trilobite: good=d13ff82 bad=def0ce3

That means the commit would have to be somewhere here:

f5a465f1a07 Promote assertion about !ReindexIsProcessingIndex to ...
57b28c08305 Doc: fix minor typos in two ECPG function descriptions.
28e858c0f95 Improve documentation and GUC description for ...
a661bf7b0f5 Remove flaky isolation tests for timeouts
874d817baa1 Multiple revisions to the GROUP BY reordering tests
466979ef031 Replace lateral references to removed rels in subqueries
a6b2a51e16d Avoid dangling-pointer problem with partitionwise ...
d360e3cc60e Fix compiler warning on typedef redeclaration
8af25652489 Introduce a new smgr bulk loading facility.
e612384fc78 Fix mistake in SQL features list
d13ff82319c Fix BF failure in commit 93db6cbda0.

My guess would be 8af25652489, as it's the only storage-related commit.

I'm currently running tests to verify this.

regards

--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company