On Fri, Feb 04, 2022 at 04:35:07PM -0500, Tom Lane wrote:
> Michael Banck <michael.banck@credativ.de> writes:
> > On Fri, Feb 04, 2022 at 02:58:59PM -0500, Tom Lane wrote:
> >> + If this seems to have affected a table, <command>REINDEX</command>
> >> + should repair the damage.
>
> > I don't think this is very helpful to the reader, are their indexes
> > corrupt or not? If we can't tell them a specific command to run to
> > check, can we at least mention that running amcheck would detect that
> > (if it actually does)? Otherwise, I guess the only way to be sure is to
> > just reindex every index? Or is this at least specific to b-trees?
>
> Yeah, I wasn't too happy with that advice either, but it seems like the
> best we can do [1]. I don't think we should advise blindly reindexing
> your whole installation, because it's a very low-probability bug.
Right ok. I wonder whether it makes sense to at hint at the low
probability then; I guess if you know Postgres well you can deduct from
the "when the last transaction that could see the tuple ends while the
page is being pruned" that it is a low-probability corner-case, but
I fear lots of users will be unable to gauge the chances they got hit by
this bug and just blindly assume they are affected (and/or ask around).
I just woke up, so I don't have any good wording suggetsions yet.
Michael
--
Michael Banck
Teamleiter PostgreSQL-Team
Projektleiter
Tel.: +49 2166 9901-171
Email: michael.banck@credativ.de
credativ GmbH, HRB Mönchengladbach 12080
USt-ID-Nummer: DE204566209
Trompeterallee 108, 41189 Mönchengladbach
Geschäftsführung: Dr. Michael Meskes, Geoff Richardson, Peter Lilley
Unser Umgang mit personenbezogenen Daten unterliegt
folgenden Bestimmungen: https://www.credativ.de/datenschutz
