On Tue, Oct 5, 2021 at 11:00 PM Alexander Lakhin <exclusion@gmail.com> wrote:
> I think that ideally pg_amcheck should not fail on a live database, that
> does not contain corrupted data, and should not affect the database
> usage by other users (as it's "only a check").
I agree that that's ideal. As you said, one or two narrow exceptions
may need to be made -- cases where there is unavoidable though weird
ambiguity (and not a report of true corruption). Overall the user
should never see failure from pg_amcheck unless the database is
corrupt, or unless things are defined in a pretty odd way, that
creates ambiguity. Ordinary DDL certainly doesn't count as unusual
here.
--
Peter Geoghegan