How can I check for metadata corruption in postgres files after a file system event like a repair guy pulling the wrong drive when replacing a failed drive (so breaking the raid set)? I found pgfsck but it claims to only support up to 8.2. Would doing a vacuum do it? pg_dump? (that takes a long time)
What I'm looking to avoid is having metadata corruption that doesn't get detected for days because the affected tuples aren't touched frequently.
Someone suggested iterating over each row in each table (e.g. select * from...). My db seems ok.
Thanks
-----Original Message----- From: m <mbkirk1024@aim.com> To: pgsql-novice <pgsql-novice@postgresql.org> Sent: Fri, Oct 29, 2010 10:23 am Subject: RE: detecting metadata corruption after dirty shutdown
>> like a repair guy pulling the wrong drive when replacing afailed drive (so breaking the raid set)
said "repair guy" being me, looking at the red light and pulling the drive next to it.