Re: pg15b2: large objects lost on upgrade

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Fri, Jul 29, 2022 at 7:16 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> I am suspicious that the problem stems from the nonstandard
>> way you've invoked psql to collect the horizon data.

> Well, I just copied the pg_dump block which occurs directly beforehand
> and modified it. I think that must take care of setting the path
> properly, else we'd have things blowing up all over the place. But the
> lack of -X could be an issue.

Hmm. Now that I look, I do see two pre-existing "naked" invocations
of psql in 002_pg_upgrade.pl, ie

$oldnode->command_ok([ 'psql', '-X', '-f', $olddumpfile, 'postgres' ],
'loaded old dump file');

$oldnode->command_ok(
[
'psql', '-X',
'-f', "$srcdir/src/bin/pg_upgrade/upgrade_adapt.sql",
'regression'
],
'ran adapt script');

Those suggest that maybe all you need is -X. However, I don't think
either of those calls is reached by the majority of buildfarm animals,
only ones that are doing cross-version-upgrade tests. So there
could be more secret sauce needed to get this to pass everywhere.

Personally I'd try to replace the two horizon-collection steps with
$newnode->psql calls, using extra_params to inject the '-o' and target
filename command line words. But if you want to try adding -X as
a quicker answer, maybe that will be enough.

regards, tom lane