Horton

Re: recovering from "found xmin ... from before relfrozenxid ..."

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Ashutosh Sharma <ashu.coek88@gmail.com>, Mark Dilger <mark.dilger@enterprisedb.com>, Masahiko Sawada <masahiko.sawada@2ndquadrant.com>, "Andrey M. Borodin" <x4mmm@yandex-team.ru>, MBeena Emerson <mbeena.emerson@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Andres Freund <andres@anarazel.de>, Peter Eisentraut <peter.eisentraut@2ndquadrant.com>, Stephen Frost <sfrost@snowman.net>
Date: 2020-09-21T14:27:30Z
Lists: pgsql-hackers 
Amit Kapila <amit.kapila16@gmail.com> writes:
> On Sun, Sep 20, 2020 at 10:43 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> AFAICS, there is no chance of the existing pg_surgery regression test
>> being fully stable if we don't fix both things.

> What if ensure that it runs with autovacuum = off and there is no
> parallel test running? I am not sure about the second part but if we
> can do that then the test will be probably stable.

Then it'll not be usable under "make installcheck", which is not
very nice.  It's also arguable that you aren't testing pg_surgery
under real-world conditions if you do it like that.

Moreover, I think that both of these points need to be addressed
anyway, as they represent bugs that are reachable independently
of pg_surgery.  Admittedly, we do not have a test case that
proves that the inconsistency between pruneheap and vacuum has
any bad effects in the absence of a7212be8b.  But do you really
want to bet that there are none?

			regards, tom lane

Commits

  1. Fix wrong data table horizon computation during backend startup.

  2. Centralize horizon determination for temp tables, fixing bug due to skew.

  3. pg_surgery: Try to stabilize regression tests.

  4. New contrib module, pg_surgery, with heap surgery functions.

  5. Set cutoff xmin more aggressively when vacuuming a temporary table.

  6. snapshot scalability: Don't compute global horizons while building snapshots.

  7. Introduce vacuum errcontext to display additional information.