Re: BUG #4897: PostgreSQL 8.3.7 crashes on Win2003 logoff

From: Gerhard Leykam <gerhard(dot)leykam(at)sealsystems(dot)de>
To: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #4897: PostgreSQL 8.3.7 crashes on Win2003 logoff
Date: 2009-07-06 13:41:11
Message-ID: 4A51FEF7.9000509@sealsystems.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Hi!

There is a known Logoff/CreateProcess problem at MS:
http://support.microsoft.com/kb/960266/en-us

But even installing this hotfix didn't solve the problem.

Nobody out there with an idea?

Bye
Gerhard Leykam

Gerhard Leykam schrieb:
> The following bug has been logged online:
>
> Bug reference: 4897
> Logged by: Gerhard Leykam
> Email address: gel123(at)sealsystems(dot)de
> PostgreSQL version: 8.3.7.1
> Operating system: Windows Server 2003 SP2
> Description: PostgreSQL 8.3.7 crashes on Win2003 logoff
> Details:
>
> Hi!
>
> That's what i have done:
>
> Installed postgresql-8.3.7-1-windows.exe on Windows Server 2003 SP2 with one
> click installer.
>
> Create some load on the machine by running example program testlibpq.exe a
> lot from a remote host. This is done by a perl script re-running
> testlibpq.exe in an endless loop started 20 times.
>
> Logoff remote desktop connection from WinSrv2003: CRASH
>
> 2009-07-01 10:15:02.515 CEST LOG: Serverprozess (PID 2512) beendete mit
> Status 128
> 2009-07-01 10:15:02.515 CEST LOG: aktive Serverprozesse werden abgebrochen
> 2009-07-01 10:15:02.734 CEST LOG: alle Serverprozesse beendet;
> initialisiere neu
> 2009-07-01 10:15:03.734 CEST FATAL: bereits bestehender Shared-Memory-Block
> wird noch benutzt
> 2009-07-01 10:15:03.734 CEST TIPP: Prüfen Sie, ob irgendwelche alten
> Serverprozesse noch laufen und beenden Sie diese.
>
> No postgres process running any more.
> Old 8.3.5 used before even can not be restarted!!! Server boot is
> necessary!
> But the 8.3.7 can be restarted and crashes again by repeating above
> scenario:
>
> 2009-07-01 10:16:36.281 CEST LOG: database system was interrupted; last
> known up at 2009-07-01 10:13:53 CEST
> 2009-07-01 10:16:36.343 CEST LOG: database system was not properly shut
> down; automatic recovery in progress
> 2009-07-01 10:16:36.453 CEST LOG: record with zero length at 0/4A9AF0
> 2009-07-01 10:16:36.453 CEST LOG: redo is not required
> 2009-07-01 10:16:36.609 CEST FATAL: the database system is starting up
> 2009-07-01 10:16:36.656 CEST LOG: Datenbanksystem ist bereit um
> Verbindungen anzunehmen
> 2009-07-01 10:16:36.703 CEST LOG: autovacuum launcher started
> 2009-07-01 10:17:35.515 CEST LOG: Serverprozess (PID 2756) beendete mit
> Status 128
> 2009-07-01 10:17:35.515 CEST LOG: aktive Serverprozesse werden abgebrochen
> 2009-07-01 10:17:35.703 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.796 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.796 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.812 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.812 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.812 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.828 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.828 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.828 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.843 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.843 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.843 CEST FATAL: the database system is in recovery
> mode
> 2009-07-01 10:17:35.953 CEST LOG: alle Serverprozesse beendet;
> initialisiere neu
> 2009-07-01 10:17:36.953 CEST FATAL: bereits bestehender Shared-Memory-Block
> wird noch benutzt
> 2009-07-01 10:17:36.953 CEST TIPP: Prüfen Sie, ob irgendwelche alten
> Serverprozesse noch laufen und beenden Sie diese.
>
> Sorry for the German/English log message mixture, but selecting English
> during installation didn't fix it?!
>
> The crash seems to be related to CreateProcess-Calls for new clients
> arriving and Windows Logoff handling. Sometimes postgres.exe reports system
> error 0xc0000142: Some ressources are missing.
>
> Bye,
> Gerhard Leykam

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Rainer Bauer 2009-07-06 13:47:12 Re: Unknown winsock error 10061
Previous Message Steve Purcell 2009-07-06 13:11:19 Re: BUG #4901: Column name "window" breaks pg_dump/pg_restore