Re: Postgresql 'eats' all mi data partition - Mailing list pgsql-bugs

From Stephan Szabo
Subject Re: Postgresql 'eats' all mi data partition
Date
Msg-id 20030925101920.A82422@megazone.bigpanda.com
Whole thread Raw
In response to Postgresql 'eats' all mi data partition  (Javier Carlos <fjcarlos@correo.insp.mx>)
Responses Re: Postgresql 'eats' all mi data partition  (Javier Carlos <fjcarlos@correo.insp.mx>)
List pgsql-bugs
On Thu, 25 Sep 2003, Javier Carlos wrote:

> ============================================================================
>                         POSTGRESQL BUG REPORT TEMPLATE
> ============================================================================
>
>
> Your name        :    Javier Carlos Rivera
> Your email address    :    fjcarlos ( at ) correo ( dot ) insp ( dot ) mx
>
>
> System Configuration
> ----------------------
>   Architecture (example: Intel Pentium)        : Intel Pentium 4
>
>   Operating System (example: Linux 2.0.26 ELF)       : Debian GNU/Linux 3.0 2.4.21
>
>   RAM                          : 256 MB
>
>   PostgreSQL version (example: PostgreSQL-6.3.2)  : PostgreSQL-7.3.4
>
>   Compiler used (example:  gcc 2.7.2)          : 2.95.4
>
>
>
> Please enter a FULL description of your problem:
> -------------------------------------------------
>     On Thursday Bruce Momjian was at Mexico; I saw him and asked about
> this problem. He told me to write to this e-mail.
>
>     When I do a simple 'UPDATE' PostgreSQL 'eats' all my partition space
> of my data directory. For example:
>
> ***** My data directory is in /var
> ***** BEFORE I do the UPDATEs I got this from df:
> OPORTUNIDADES:~# df
> Filesystem           1k-blocks      Used Available Use% Mounted on
> /dev/hda2              2885812     14372   2724848   1% /
> /dev/hda1             14421344   1195132  12493652   9% /var
> /dev/hda3              7692908    888560   6413568  13% /usr
> /dev/hda6             12491804     22704  11834536   1% /javier
> /dev/hda7              1494204     23936   1394364   2% /home
>
>
> ***** Then I do the UPDATEs:
> **** The updates are of this type :
> UPDATE tbl_personas SET "PIDFOLIO"=-2 WHERE "PIDFOLIO" IS NULL;
> UPDATE tbl_personas SET "P5_1"=-2 WHERE "P5_1" IS NULL;
> UPDATE tbl_personas SET "P4PAQ"=-2 WHERE "P4PAQ" IS NULL;
> UPDATE tbl_personas SET "P5_4"=-2 WHERE "P5_4" IS NULL;
> UPDATE tbl_personas SET "P5_5"=-2 WHERE "P5_5" IS NULL;
> UPDATE tbl_personas SET "P36_4"=-2 WHERE "P36_4" IS NULL;
> ..
> UPDATE table_name SET column_name = -2 WHERE column_name IS NULL;
> ..

If you're not vacuuming, you're going to have a potentially large
number of dead rows.  Does a vacuum between updates or a vacuum full at
the end bring the space usage down to something reasonable?

pgsql-bugs by date:

Previous
From: Andreas Pflug
Date:
Subject: Re: Postgresql 'eats' all mi data partition
Next
From: Stephan Szabo
Date:
Subject: Re: Postgresql 'eats' all mi data partition