Re: slow queries over information schema.tables

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Thu, Dec 6, 2018 at 11:32 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> It's fairly hard to imagine practical cases where we'd not call
>> AcceptInvalidationMessages at least once per query, so I'm not
>> very sure what you're on about.

> Unless I'm confused, it happens any time you run a query that only
> touches tables using lockmodes previously acquired by the current
> transaction. Like:

> BEGIN;
> some query;
> the same query again;

In my testing, that still hits AIM() during parserOpenTable().

[ further experimentation... ] It looks like if you prepare
a query and then just execute it repeatedly in one transaction,
you'd not reach AIM (as long as you were getting generic plans).
Possibly that's a gap worth closing.

regards, tom lane