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

trying to run PITR recovery

From: Warren Little <Warren(dot)Little(at)MeridiasCapital(dot)com>
To: pgsql-admin(at)postgresql(dot)org
Subject: trying to run PITR recovery
Date: 2007-03-23 23:16:08
Message-ID: 63EA6DBE-34B9-4D2D-A61C-D7EF38B3D9DB@MeridiasCapital.com (view raw or flat)
Thread:
Lists: pgsql-admin
Hello,
I'm testing my PITR recovery procedures and something doesn't look  
right.
The following is from the logs upon starting postgres with  
recovery.conf file in place

2007-03-23 05:56:00 MDTLOG:  database system was interrupted at  
2007-03-18 05:09:15 MDT
@ 2007-03-23 05:56:00 MDTLOG:  starting archive recovery
@ 2007-03-23 05:56:00 MDTLOG:  restore_command = "cp /data/pgLocal/ 
archive/WAL_restore/%f "%p""
cp: cannot stat `/data/pgLocal/archive/WAL_restore/00000001.history':  
No such file or directory
@ 2007-03-23 05:56:00 MDTLOG:  restored log file  
"000000010000011A000000EE.004E0060.backup" from archive
@ 2007-03-23 05:56:00 MDTLOG:  restored log file  
"000000010000011A000000EE" from archive
@ 2007-03-23 05:56:00 MDTLOG:  checkpoint record is at 11A/EE4E0060
@ 2007-03-23 05:56:00 MDTLOG:  redo record is at 11A/EE4E0060; undo  
record is at 0/0; shutdown FALSE
@ 2007-03-23 05:56:00 MDTLOG:  next transaction ID: 2415965426; next  
OID: 81701223
@ 2007-03-23 05:56:00 MDTLOG:  next MultiXactId: 43380; next  
MultiXactOffset: 92368
@ 2007-03-23 05:56:00 MDTLOG:  automatic recovery in progress
@ 2007-03-23 05:56:01 MDTLOG:  redo starts at 11A/EE4E00B0
@ 2007-03-23 05:56:15 MDTLOG:  restored log file  
"000000010000011A000000EF" from archive
@ 2007-03-23 05:56:27 MDTLOG:  restored log file  
"000000010000011A000000F0" from archive
@ 2007-03-23 05:56:31 MDTLOG:  restored log file  
"000000010000011A000000F1" from archive
@ 2007-03-23 05:56:34 MDTLOG:  restored log file  
"000000010000011A000000F2" from archive
@ 2007-03-23 05:56:40 MDTLOG:  restored log file  
"000000010000011A000000F3" from archive
@ 2007-03-23 05:56:50 MDTLOG:  restored log file  
"000000010000011A000000F4" from archive
@ 2007-03-23 05:57:02 MDTLOG:  restored log file  
"000000010000011A000000F5" from archive
@ 2007-03-23 05:57:07 MDTLOG:  restored log file  
"000000010000011A000000F6" from archive
@ 2007-03-23 05:57:11 MDTLOG:  restored log file  
"000000010000011A000000F7" from archive
@ 2007-03-23 05:57:14 MDTLOG:  restored log file  
"000000010000011A000000F8" from archive
@ 2007-03-23 05:57:18 MDTLOG:  restored log file  
"000000010000011A000000F9" from archive
@ 2007-03-23 05:57:21 MDTLOG:  restored log file  
"000000010000011A000000FA" from archive
@ 2007-03-23 05:57:27 MDTLOG:  restored log file  
"000000010000011A000000FB" from archive
@ 2007-03-23 05:57:29 MDTLOG:  restored log file  
"000000010000011A000000FC" from archive
@ 2007-03-23 05:57:33 MDTLOG:  restored log file  
"000000010000011A000000FD" from archive
@ 2007-03-23 05:57:35 MDTLOG:  incorrect resource manager data  
checksum in record at 11A/FD492B20
@ 2007-03-23 05:57:35 MDTLOG:  redo done at 11A/FD492210
@ 2007-03-23 05:57:36 MDTLOG:  restored log file  
"000000010000011A000000FD" from archive
@ 2007-03-23 05:57:36 MDTLOG:  archive recovery complete
@ 2007-03-23 05:57:36 MDTLOG:  could not truncate directory  
"pg_multixact/members": apparent wraparound
@ 2007-03-23 05:59:00 MDTLOG:  database system is ready
@ 2007-03-23 05:59:00 MDTLOG:  transaction ID wrap limit is  
3065701724, limited by database "postgres"
@ 2007-03-23 12:14:04 MDTLOG:  autovacuum: processing database  
"postgres"

My concern is that there were many more logfiles to be played  
following "00000010000011A000000FD"
(ie 000000010000011E0000005C) yet it appears the recovery stop at  
that point and called it good.
I would assume all WAL logs would be restored.

The recovery.conf was simple:
restore_command = 'cp /data/pgLocal/archive/WAL_restore/%f "%p"'

The backup file 000000010000011A000000EE.004E0060.backup" contained:
START WAL LOCATION: 11A/EE4E0060 (file 000000010000011A000000EE)
STOP WAL LOCATION: 11A/EFF68AB8 (file 000000010000011A000000EF)
CHECKPOINT LOCATION: 11A/EE4E0060
START TIME: 2007-03-17 20:29:16 MDT
LABEL: 076_pgdata.tar
STOP TIME: 2007-03-18 05:16:17 MDT

Does the line:  incorrect resource manager data checksum in record at  
11A/FD492B20
mean there is a corrupted WAL log file?

Any insight here would be helpful
version PG 8.1.2 64 bit Linux


thanks


Warren Little
Chief Technology Officer
Meridias Capital Inc
ph 866.369.7763



Responses

pgsql-admin by date

Next:From: Daniel Ricardo MedinaDate: 2007-03-23 23:19:03
Subject: Re: no verification of client certificate?
Previous:From: Peter KoczanDate: 2007-03-23 22:03:22
Subject: Postgres Perl Module and Kerberos problem on Solaris

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