Marko Tiikkaja <marko.tiikkaja@cs.helsinki.fi> writes:
> Another thing I now see is this:
> BEGIN;
> SELECT pg_advisory_xact_lock(1);
> -- do something here
> -- upgrade to session lock
> SELECT pg_advisory_lock(1);
> COMMIT;
> This seems useful, since the xact lock would be automatically released
> if an error happens during "-- do something here" so you wouldn't need
> to worry about releasing the lock elsewhere. But I'm not sure this is
> safe. Can anyone see a problem with it?
I think the POLA dictates that the behavior of that should be that you
now have both a transactional and a nontransactional hold on the lock;
and only the transactional hold goes away at commit.
regards, tom lane