Re: BUG #4096: PG 8.3.1. confused about remaining disk space - Mailing list pgsql-bugs

From J6M
Subject Re: BUG #4096: PG 8.3.1. confused about remaining disk space
Date
Msg-id 009f01c89af6$2aaa1100$1500020a@eden.fr
Whole thread Raw
In response to BUG #4096: PG 8.3.1. confused about remaining disk space  ("" <j6m@adm.estp.fr>)
List pgsql-bugs
After further investigations, it seems the second hypothesis is correct.
This particular DB was created with "with OIDs" as default.

The script I was using had no SET for this option and it creates a few
temporary tables (which are quite small in size as they are created by
SELECT .. INTO TEMP TABLE .. FROM .. GROUP BY .., so they are smaller than
their origins).

Adding SET default_with_oids = true to the script seems to solve the
problem.

Also, when run with the former releases, AUTOVACUUMING was off. Now it is
on.

Regards
J6M
----- Original Message -----
From: "J6M" <j6m@adm.estp.fr>
To: "Stefan Kaltenbrunner" <stefan@kaltenbrunner.cc>
Cc: "Tom Lane" <tgl@sss.pgh.pa.us>; <pgsql-bugs@postgresql.org>
Sent: Tuesday, April 08, 2008 12:37 PM
Subject: Re: [BUGS] BUG #4096: PG 8.3.1. confused about remaining disk space


> The df was made while the error messages occured and the scripts were
> running. So it is not a temporary 'mammothean' table.
>
> When I thought of this, two possible elads came to my mind :
>
> - A mistake when I tuned the postgresql.conf file (resource is scarce, as
> this test machine is a PIII 700 with 256 MB RAM dating from Q4 2000)
>
> - The default for this particular DB has been set to keep OIDs (I "pour"
> several INSERTs sets arising from different files in an auxiliary table,
> and keep the last entered entry for any relevant keys, Those entries then
> go to a production table).
>

pgsql-bugs by date:

Previous
From: Michael Meskes
Date:
Subject: Re: BUG #4091: ECPG $select incompability with Informix
Next
From: "Prakash Suryawanshi"
Date:
Subject: Error in module