Re: oldest xmin is far in the past

From: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
To: John Snow <sleepwalker(dot)js(at)gmail(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: oldest xmin is far in the past
Date: 2016-03-19 16:32:59
Message-ID: 16f29a07-1db8-9188-7202-d834466fe7ac@2ndquadrant.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On 03/19/2016 06:29 AM, John Snow wrote:
> There is no any long transaction neither prepared transaction.

Can you show us pg_stat_activity? Particularly the xmin values for
backends attached to the two databases mentioned in the log (1 and 12451).

FWIW the second OID is a bit weird - the first OID assigned to normal
objects is defined as 16384, and none of the so I wonder how you managed
to create a database with such DB?

Unless it's one of the template databases, but I got different OIDs when
I tried a fresh initdb on 9.4.

> #autovacuum_freeze_max_age = 200000000 - default value

After looking at the code a bit more, I see it uses some additional
configuration options:

* freeze_min_age
* vacuum_freeze_min_age
* autovacuum_freeze_max_age (we already know this one)

What values are set for those?

regards

--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Fabien COELHO 2016-03-19 16:40:44 Re: extend pgbench expressions with functions
Previous Message Fabien COELHO 2016-03-19 16:08:20 Re: pgbench stats per script & other stuff