BUG #6484: pgstat wait timeout

[cneo]

The following bug has been logged on the website:

Bug reference:      6484
Logged by:          Charlie Neo
Email address:      [hidden email]
PostgreSQL version: 9.0.6
Operating system:   Windows Server 2008 R2 Enterprise
Description:        

During heavy database loading, "pgstat wait timeout" is showing up in the
logs. When this happens pgstat.stat file stop updating. Stats in
pg_stat_user_tables also stop updating. Resulting in autovacuum not
triggering and greatly degrades performs. Killing the stats collection
process causing it to re-spawn does not help, as it will quickly go back
into the same state without any new stats collected.

Stack trace of the process postgres.exe -forkcol

ntoskrnl.exe!memset+0x64a
ntoskrnl.exe!KeWaitForMultipleObjects+0xf5d
ntoskrnl.exe!KeWaitForMultipleObjects+0x26a
ntoskrnl.exe!NtWaitForSingleObject+0x40f
ntoskrnl.exe!NtWaitForSingleObject+0x77e
ntoskrnl.exe!KeSynchronizeExecution+0x3a43
ntdll.dll!ZwWaitForMultipleObjects+0xa
KERNELBASE.dll!GetCurrentProcess+0x40
kernel32.dll!WaitForMultipleObjectsEx+0xb3
postgres.exe!reltimesend+0x1f8d50
postgres.exe!reltimesend+0x1f9053
postgres.exe!reltimesend+0x2054af
postgres.exe!reltimesend+0x20aa85
postgres.exe!reltimesend+0x14af4e
postgres.exe!reltimesend+0x38a65d
kernel32.dll!BaseThreadInitThunk+0xd
ntdll.dll!RtlUserThreadStart+0x21


--
Sent via pgsql-bugs mailing list ([hidden email])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-bugs

[Robert Haas]

On Wed, Feb 22, 2012 at 6:04 PM,  <[hidden email]> wrote:

> The following bug has been logged on the website:
>
> Bug reference:      6484
> Logged by:          Charlie Neo
> Email address:      [hidden email]
> PostgreSQL version: 9.0.6
> Operating system:   Windows Server 2008 R2 Enterprise
> Description:
>
> During heavy database loading, "pgstat wait timeout" is showing up in the
> logs. When this happens pgstat.stat file stop updating. Stats in
> pg_stat_user_tables also stop updating. Resulting in autovacuum not
> triggering and greatly degrades performs. Killing the stats collection
> process causing it to re-spawn does not help, as it will quickly go back
> into the same state without any new stats collected.
>
> Stack trace of the process postgres.exe -forkcol
>
> ntoskrnl.exe!memset+0x64a
> ntoskrnl.exe!KeWaitForMultipleObjects+0xf5d
> ntoskrnl.exe!KeWaitForMultipleObjects+0x26a
> ntoskrnl.exe!NtWaitForSingleObject+0x40f
> ntoskrnl.exe!NtWaitForSingleObject+0x77e
> ntoskrnl.exe!KeSynchronizeExecution+0x3a43
> ntdll.dll!ZwWaitForMultipleObjects+0xa
> KERNELBASE.dll!GetCurrentProcess+0x40
> kernel32.dll!WaitForMultipleObjectsEx+0xb3
> postgres.exe!reltimesend+0x1f8d50
> postgres.exe!reltimesend+0x1f9053
> postgres.exe!reltimesend+0x2054af
> postgres.exe!reltimesend+0x20aa85
> postgres.exe!reltimesend+0x14af4e
> postgres.exe!reltimesend+0x38a65d
> kernel32.dll!BaseThreadInitThunk+0xd
> ntdll.dll!RtlUserThreadStart+0x21

This is a known issue, but I am not sure anyone has a plan for what to
do about it.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

--
Sent via pgsql-bugs mailing list ([hidden email])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-bugs