I am new to postgres administration - been a user for a long time. I inherited about 300 machines on a client's network that are linux based (2.4-18 and EXT3) running postgresql 7.2.1 and see random disk corruption pretty consistently - about one a week.

The postgresql.conf looks pretty plain vanilla so lots of opportunities for some tuning but that is a different post. All I have done so far is bump shhmax - big difference.

Similar machines with the same kernel that are running mysql have not had any failures so I am looking at the db. Just wondering if anyone else has seen this problem and what the best way to pinpoint where it is coming from would be.

thanks ...