Failing to recover after panic shutdown

Hi

I am running Postgres 8.2 on Windows 2003 server SP2.

Every now and then (2-3 times a year) our Postgres service is down and we need to manually start it. This is what we find:

In log when going down:
2008-06-02 13:40:02 PANIC: could not open file "pg_xlog/000000010000001C00000081" (log file 28, segment 129): Invalid argument

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
2008-06-02 13:40:02 LOG: server process (PID 5792) exited with exit code 3
2008-06-02 13:40:02 LOG: terminating any other active server processes
2008-06-02 13:40:02 WARNING: terminating connection because of crash of another server process
2008-06-02 13:40:02 DETAIL: The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2008-06-02 13:40:02 HINT: In a moment you should be able to reconnect to the database and repeat your command.
(WARNING, DETAIL and HINT is repeated a few times...)
2008-06-02 13:40:02 LOG: all server processes terminated; reinitializing
2008-06-02 13:40:02 LOG: database system was interrupted at 2008-06-02 13:39:39 W. Europe Daylight Time
2008-06-02 13:40:02 LOG: checkpoint record is at 1C/80F646B0
2008-06-02 13:40:02 LOG: redo record is at 1C/80F646B0; undo record is at 0/0; shutdown FALSE
2008-06-02 13:40:02 LOG: next transaction ID: 0/316291661; next OID: 44447343
2008-06-02 13:40:02 LOG: next MultiXactId: 1929; next MultiXactOffset: 4093
2008-06-02 13:40:02 LOG: database system was not properly shut down; automatic recovery in progress
2008-06-02 13:40:02 LOG: redo starts at 1C/80F64700
2008-06-02 13:40:02 LOG: unexpected pageaddr 1C/79000000 in log file 28, segment 129, offset 0
2008-06-02 13:40:02 LOG: redo done at 1C/80FFD6E0
2008-06-02 13:40:02 FATAL: the database system is starting up
2008-06-02 13:40:03 FATAL: the database system is starting up
2008-06-02 13:40:03 FATAL: the database system is starting up
2008-06-02 13:40:03 FATAL: the database system is starting up
2008-06-02 13:40:03 FATAL: could not open file "pg_xlog/000000010000001C00000081" (log file 28, segment 129): Invalid argument
2008-06-02 13:40:03 FATAL: the database system is starting up
2008-06-02 13:40:03 LOG: startup process (PID 4420) exited with exit code 1
2008-06-02 13:40:03 LOG: aborting startup due to startup process failure
2008-06-02 13:40:05 LOG: archived transaction log file "000000010000001C00000080"
2008-06-02 13:40:05 LOG: logger shutting down

In log when manually started:

2008-06-02 13:50:34 LOG: database system was shut down at 2008-06-02 13:40:03 W. Europe Daylight Time
2008-06-02 13:50:34 LOG: checkpoint record is at 1C/80FFE990
2008-06-02 13:50:34 LOG: redo record is at 1C/80FFE990; undo record is at 0/0; shutdown TRUE
2008-06-02 13:50:34 LOG: next transaction ID: 0/316292191; next OID: 44447343
2008-06-02 13:50:34 LOG: next MultiXactId: 1929; next MultiXactOffset: 4093
2008-06-02 13:50:34 LOG: database system is ready
2008-06-02 13:50:35 LOG: archived transaction log file "000000010000001C00000080"
2008-06-02 13:50:54 LOG: archived transaction log file "000000010000001C00000080"

We are archiving WAL-segments at a remote machine, and we are copying non-filled WAL-segments every 10 minutes to be able to rebuild the DB with a maximum of 10 minutes of missing data. (I don't know if that has anything to do with it).

Does anyone have a clue on what the problem could be?

Vennlig hilsen
Per Lauvås
Systemutvikler
Fax: +47 22 41 60 61
Direct: +47 24 15 55 51
Mintra as
Storgata 1
P.O. Box 8945 Youngstorget
N-0028 Oslo
Tel: +47 24 15 55 00
Fax: +47 22 41 60 61
http://www.mintra.no
Faglig forum: Helhetlig læring og kompetanse
Motta Mintras faglige e-postmagasin Helhetlig LÆRING & KOMPETANSE - gratis hver måned. Svar på denne e-posten med "HLK" i emnefeltet eller se www.mintra.no for å melde din interesse.

Hyllevare e-læringskurs på Trainingportal.no
Få tilgang til Mintras hyllevarer av e-læringskurs innen HMS, applikasjonsopplæring, prosjektledelse, teknisk opplæring med mer på www.trainingportal.no