The checksum error only was reported only once so I guess it was automatically fixed immediately.
The log was output only one time while the box was running very long time (> 1 year.).
I have hundreds of the machines with same configuration running very long time as well,
but the incident was only once.
Can we point out a specific bug that can lead to this?
Or HW issue is the most probable cause?
I don't see any HW error logs.
Regards,
Tatsuki
From: pgsql-general-owner@postgresql.org <pgsql-general-owner@postgresql.org> on behalf of John R Pierce <pierce@hogranch.com> Sent: Monday, August 22, 2016 12:42:41 PM To: pgsql-general@postgresql.org Subject: Re: [GENERAL] incorrect checksum detected on "global/pg_filenode.map" when VACUUM FULL is executed
On 8/21/2016 8:37 PM, Michael Paquier wrote:
PosgreSQL version is 9.2.4.
You are missing a couple of years worth of bug fixes here..
3+ years, to be more specific, 9.2.4 was released in april 2013.
indeed, and several of the bug fixes involved data corruption. current is 9.2.18