Bertrand Drouvot <bertranddrouvot...@gmail.com> writes:
> On Fri, Jan 24, 2025 at 11:42:15AM -0500, Tom Lane wrote:
>> PS: FTR, the hits I got on this in the past 90 days were
>>
>> sysname | branch | snapshot | stage |
>> l
>>
>> ---------+---------------+---------------------+---------------+------------------------------------------------------------------------------------------------------------
>> drongo | HEAD | 2024-10-29 10:52:07 | recoveryCheck | # Failed
>> test 'activeslot slot invalidation is logged with vacuum on pg_class'\r\r
>> drongo | REL_16_STABLE | 2024-10-31 08:07:11 | recoveryCheck | # Failed
>> test 'activeslot slot invalidation is logged with vacuum on pg_authid'\r\r
>> ...
> Out of curiosity, how did you generate this output? (that looks wery useful)
That is from a SQL query on the database that stores the last few
months' worth of buildfarm logs. I believe current policy is that
only committers will be granted access to that machine, partly
because of security concerns and partly because of resource
limitations. But in the interests of transparency, what I did was
pgbfprod=> select * from
(select sysname, s.branch, snapshot, stage,
unnest(string_to_array(log_text, E'\n')) as l
from build_status s join build_status_log l using (sysname, snapshot)
where stage != 'OK' and snapshot > localtimestamp - '3 mons'::interval) ss
where l like '%Failed test%activeslot slot invalidation%'
\g outfile
(Other than the specific string-to-search-for, this is a canned
query pattern that I use a lot. It scans all failed tests' logs
in the past 3 months.) I did some manual duplicate-removal
and hand filtering afterwards, IIRC.
regards, tom lane
