Giorgio Saviane <gsaviane@gmail.com> writes:
> Ok, let me add one more step in that sequence.
> Before shutting down Postgres and going down to single mode we stopped the
> application server and other services around it. I'm pretty sure that there
> was a reasonable window of time before we went to single mode.
Hmm, but did you stop the connection pooler? That's what was holding
the database session open, if I'm understanding things correctly.
Unless you've changed the default autovacuum settings (particularly
autovacuum_naptime), I'd have expected autovacuum to start zapping
temp tables within a minute or so after they become orphaned.
Even more to the point, if the database session was allowed to stop
normally, it should've cleaned them up itself.
regards, tom lane