Re: Template0 datfrozenxid age is 160million and progressing

From: Stephen Eilert <contact(at)stepheneilert(dot)com>
To: pgsql-general(at)lists(dot)postgresql(dot)org
Subject: Re: Template0 datfrozenxid age is 160million and progressing
Date: 2018-08-01 20:08:37
Message-ID: 1533154117.3567315.1460357192.6B184715@webmail.messagingengine.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

160 million is a very low number. I manage production databases which
reach this value in a day easily. As other said, 200 million is the
default threshold for the anti-wraparound vacuums. I wouldn't worry,
specially for template0.
That said, there is nothing preventing you from temporarily changing
DATALLOWCONN, running vaccuum (which should be very quick) and then
changing it back. But you should not have to. I do that in our
production database, but only because it suffers from a bad schema
design and we ended up with thousands of tables, which is too much for
the autovacuum workers to cope alone, so they need a manual "boost". I
still don't disable autovacuum.
I don't really understand the bit about autovacuum changing query
performance. In which scenario would it be preferable to have outdated
analyzer statistics? This would be like running a system with garbage
collection disabled because GC can increase the amount of free memory.
That's the whole point.
— Stephen

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Richard Schmidt 2018-08-01 21:09:30 Pg_rewind cannot load history wal
Previous Message Adrian Klaver 2018-08-01 18:15:13 Re: Adding terminal title support for psqlrc