Yes, most of the time PostgreSQL only complains about missing pg_clog/
files then complains about invalid page headers. The "could not create
... File exists" message was only seen a couple of times.
I don't have a system that exhibits this problem right now. I can check
the XID counter when it happens again.
Thanks!
-----Original Message-----
From: Tom Lane [mailto:tgl@sss.pgh.pa.us]=20
Sent: Friday, August 24, 2007 12:11 PM
To: Stefan Kaltenbrunner
Cc: Feng Chen; Alvaro Herrera; Gregory Stark; Marc Schablewski;
Decibel!; pgsql-bugs@postgresql.org
Subject: [SPAM] Re: [SPAM] Re: [BUGS] BUG #3484: Missing pg_clog file /
corrupt index=20
Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes:
> Feng Chen wrote:
>> VERSION =3D PostgreSQL 8.1.2
> this could be
> http://archives.postgresql.org/pgsql-committers/2006-01/msg00287.php
Note, however, that that only explains the "could not create ... File
exists" complaint, which is hardly the main problem here.
The combination of "invalid page header" and "could not access status of
transaction" messages looks like nothing so much as severe data
corruption. Remember that the xmin/xmax fields are basically the first
thing we can check with any degree of strictness when examining a
tuple. This means that if a page is partially clobbered, but not in a
way that sets off the invalid-page-header checks, then the odds are very
high that the first detectable sign of trouble will be references to
transaction numbers that are far away from what the system is really
using. (Is 1481866610 anywhere near the current XID counter reported
by pg_controldata?)
I concur with the upthread comment to check into kernel bugs,
particularly if all of your machines are using the same old kernel
release.
regards, tom lane