On 11 April 2018 at 03:14, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> 2. Do we want to revert Andrew's test stabilization patch? If I
> understand correctly, the problem is the inverse of what was diagnosed:
> "any running transaction at the time of the test could prevent pages
> from being set as all-visible". That's correct, but the test doesn't
> depend on pages being all-visible -- quite the contrary, it depends on
> the pages NOT being all-visible (which is why the HeapFetches counts are
> all non-zero). Since the pages contain very few tuples, autovacuum
> should never process the tables anyway.
I think it's probably a good idea to revert it once the
instrumentation is working correctly. It appears this found a bug in
that code, so is probably useful to keep just in case something else
breaks it in the future.
I don't think there is too much risk of instability from other
sources. There's no reason an auto-vacuum would trigger and cause a
change in heap fetches. We only delete one row from lprt_a, that's not
going to trigger an auto-vacuum.
--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services