Re: template0 needing vacuum freeze?

From: Don Seiler <don(at)seiler(dot)us>
To: Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at>
Cc: Postgres General <pgsql-general(at)postgresql(dot)org>
Subject: Re: template0 needing vacuum freeze?
Date: 2020-05-19 01:45:07
Message-ID: CAHJZqBBDOmhg488UaCNwzDdoLDxD3YHOF_AkUHVyLmHA4gGeLg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Mon, May 18, 2020 at 1:40 AM Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at>
wrote:

>
> Did you see any weird messages when you vacuumed "template0"?
>

No.

> Did "datfrozenxid" shrink after the operation?
>
> "foo_db" seems to be the bigger problem.

Perhaps autovacuum never handled "template0" because it concluded (rightly)
> that
> it has to deal with "foo_db" first.
>

Yes this DB had a table in it that had been autovacuuming since Feb 2. It's
age is half way to wraparound so I'm in the middle of a manual VACUUM
FREEZE on it. I'd be interested in knowing if that prevents template0 from
autovacuuming itself. There are no other autovacuum jobs running.

>
> > I can say that these DB
> > has previously been altered for locale changes as well.
>
> Would you care to explain that? You changed "template0"? How?
>

It was changed before my time here to change the encoding from LATIN1 to
UTF manually. One of the fun treats I'm working to correct as I also
prepare these for upgrade to PG12.

What are your non-default autovacuum settings? Perhaps you should speed up
> autovacuum
> by reducing "autovacuum_vacuum_cost_delay" to 2ms or less, and by
> increasing
> "maintenance_work_mem".
>

All autovacuum settings on this DB are default. Cost delay is at the
default 20ms. maintenance_work_mem I've already increased to 512MB (this VM
has 8GB RAM).

--
Don Seiler
www.seiler.us

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Tom Lane 2020-05-19 01:51:11 Re: template0 needing vacuum freeze?
Previous Message Albrecht Dreß 2020-05-18 16:43:14 Unique index on hash of jsonb value - correct solution?