Re: Vacuum stops with misleading max_fsm_pages error - Mailing list pgsql-admin

From Kieran Cooper, Lyris UK
Subject Re: Vacuum stops with misleading max_fsm_pages error
Date
Msg-id 00a801c780c3$0aad0e00$0b02a8c0@KieranX15
Whole thread Raw
In response to Vacuum stops with misleading max_fsm_pages error  ("Kieran Cooper, Lyris UK" <kieran@lyris.co.uk>)
List pgsql-admin
> "Joshua D. Drake" <jd@commandprompt.com> writes:
>> Kieran Cooper, Lyris UK wrote:
>>> INFO:  free space map contains 20914 pages in 61 relations
>>> DETAIL:  A total of 14992 page slots are in use (including overhead).
>>> 14992 page slots are required to track all free space.
>>> Current limits are:  900000 page slots, 6000 relations, using 5659 kB.
>
>> I am not sure what your question is. The above looks perfectly
>> reasonable.

> I think he's wondering why the second number is less than the first.
> AFAICT that should be impossible after a VACUUM FULL, but there are
> probably tables that haven't been touched by the VACUUM FULL --- stuff
> in other databases being one obvious possibility.  As for the vacuum
> not having done every table in the current database, did you run it
> as superuser?
>
> regards, tom lane

Thanks for your messages Tom and Josh.

The problem I'm finding is that the Vacuum finishes before it has touched
all the tables in the database. When I run it on the same database in the
same way on other servers, it does all the tables. There is only 1 database
on this machine (in addition to template1 and template2). I'm running it
from within the postgres command line, logged in as postgres.

Thanks
Kieran



pgsql-admin by date:

Previous
From: "Matthew T. O'Connor"
Date:
Subject: Re: Vacuum stops with misleading max_fsm_pages error
Next
From: "Kieran Cooper, Lyris UK"
Date:
Subject: Re: Vacuum stops with misleading max_fsm_pages error