Skip site navigation (1) Skip section navigation (2)

Re: "global/pgstat.stat" corrupt

From: David Kerr <dmk(at)mr-paradox(dot)net>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: "global/pgstat.stat" corrupt
Date: 2012-06-22 20:11:26
Message-ID: 20120622201126.GC98691@mr-paradox.net (view raw or flat)
Thread:
Lists: pgsql-performance
On Fri, Jun 22, 2012 at 03:49:00PM -0400, Tom Lane wrote:
- David Kerr <dmk(at)mr-paradox(dot)net> writes:
- > I just restored a DB from a cold backup (pg_ctl stop -m fast)
- 
- > When starting the DB I see:
- > LOG:  corrupted statistics file "global/pgstat.stat"
- 
- Is that repeatable?  It wouldn't be too surprising to see this once when
- starting from a filesystem backup, if you'd managed to capture a
- partially written stats file in the backup.

hmm, possibly. it's somewhat worrysome if that file isn't written out
what else isn't getting written out.

our backups are SAN based snapshots, maybe i need to "sync" prior to the snapshot
even with the DB being down.

- > When I look at the filesystem I don't see a global/pgstat.stat file but i do see a 
- > pg_stat_tmp/pgstat.stat
- 
- This is normal --- global/pgstat.stat is only supposed to exist when the
- system is shut down.  Transient updates are written into pg_stat_tmp/

ah ok

- > Are those stats related to pg_stats style statistics?
- 
- No, this file is for the stats collection subsystem,
- http://www.postgresql.org/docs/9.1/static/monitoring-stats.html
- 
- 			regards, tom lane
- 

Oh, hmm thanks. I had looked at that before but must not have fully understood
what was going on. That's interesting!

Dave

In response to

pgsql-performance by date

Next:From: C├ędric VillemainDate: 2012-06-25 15:22:00
Subject: Re: index-only scan is missing the INCLUDE feature
Previous:From: Tom LaneDate: 2012-06-22 19:49:00
Subject: Re: "global/pgstat.stat" corrupt

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group