table corruption... :(

From: "Thilo Hille" <thilo(at)resourcery(dot)de>
To: <pgsql-novice(at)postgresql(dot)org>
Subject: table corruption... :(
Date: 2002-09-20 11:35:11
Message-ID: 006901c26099$c7a160a0$0b00a8c0@resourcery.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-novice

hi,
today i had 2 corrupted tables after nearly 2 month without any problems.
i excluded the damaged records (1 in each table) and rebuild both tables.
one of the tables is used as some kind of "first in last out" queue to store
continous data for a time slot about 1 day which is about 80.000 records, so
the content of the table changes daily.
another table queues data for 3 month which is about 2.000.000 records.
currently i do a vacuum at night to keep the database clean and accesstimes
usable. i dont have a clue of where the corruption came from.
is it mandatory for the tables integrity to run "analyze" periodically?
the sytem is a AMD Athlon, 1 gb memory running Red Hat Linux 7.1
pg version is 7.2-1PGDG

excerpt from postgresql.conf:
shared_buffers =35000
max_fsm_relations = 100
max_fsm_pages = 2000
sort_mem = 128
vacuum_mem = 8192

clients:
php-4.1.2
perl v5.6.1 with DBI::version "1.18"

any ideas how to track down the cause for corruptions like this?
any problems known with the pg version?

thanks thilo

In response to

Responses

Browse pgsql-novice by date

  From Date Subject
Next Message Denny Permana 2002-09-20 16:08:16 Unsubscribe Me Please
Previous Message Rory Campbell-Lange 2002-09-19 17:09:35 Re: Make an id field max(id)+1 rather than SERIAL