Re: Improving tracking/processing of buildfarm test failures

Hello hackers,

Please take a look at the September report on buildfarm failures:

# SELECT br, COUNT(*) FROM failures WHERE dt >= '2026-09-01' AND
 dt < '2026-10-01' GROUP BY br;
REL_14_STABLE: 1
REL_15_STABLE: 7
REL_16_STABLE: 4
REL_17_STABLE: 8
REL_18_STABLE: 9
REL_19_STABLE: 73
master: 48
-- Total: 150
(Counting test failures only, excluding indent-check, Configure, Build
errors, also excluding failures from fruitcrow.)

# SELECT COUNT(*) FROM (SELECT DISTINCT issue_link FROM failures WHERE
 dt >= '2026-09-01' AND dt < '2026-10-01');
26

# SELECT issue_link, COUNT(*) FROM failures WHERE dt >= '2026-09-01' AND
 dt < '2026-10-01' GROUP BY issue_link ORDER BY 2 DESC LIMIT 5;
https://www.postgresql.org/message-id/235092.1789870332%40sss.pgh.pa.us : 27
-- XversionUpgrade-REL9_2_STABLE-REL_19_STABLE fails due to missing datminmxid column (after 5cd84ed62)
-- Fixed

https://www.postgresql.org/message-id/d92b18e8-00aa-4c32-95ff-7d807a427481%40vondra.me : 18
-- occasional ECPG failures on dikkop (FreeBSD) due to non-thread-safe snprintf

https://www.postgresql.org/message-id/CA%2BHiwqGitHCmO7nvV4nXXYWDkNYA%2BzVcJj%3DJNmL3C1O9B0Gz0A%40mail.gmail.com : 14
-- windows.sql and equivclass.sql fail due to plan changes (after c62b33091)
-- Fixed

https://www.postgresql.org/message-id/rpkhzrrmo7qiz7rr3qlufksb2oci33wcg3ao75le7dqindgoif%406yozkjwb6bdf : 11
-- 001_late_shmem_alloc.pl fails with huge pages configured (afer 07fdee7c8)
-- Fixed

https://www.postgresql.org/message-id/511b76ca-8aff-4f6a-b1e5-18d1467affb0%40gmail.com : 8
-- test_decoding/ddl.sql failed due to assert in ReorderBufferBuildTupleCidHash() triggered

# SELECT COUNT(*) FROM failures WHERE dt >= '2026-09-01' AND
 dt < '2026-10-01' AND issue_link IS NULL; -- Unsorted/unhelpful failures
24

Short-lived failures: 59

Best regards,
Alexander