Обсуждение: Hanging queries on Windows 2003 SP1

Hi,

We are seeing hanging queries on Windows 2003 Server SP1 with dual CPU,
looks like one of the process is blocked. In a lot of cases, the whole
DB is blocked if this process is holding important locks.

Looks like this issue was discussed in the following thread a few month
ago, but didn't seem to have a solution mention. I would liek to know
if there is a patch for this already?

http://archives.postgresql.org/pgsql-performance/2006-03/msg00129.php

I would appreciate your feedbaek,

Thanks,
Wei

> Hi,
>
> We are seeing hanging queries on Windows 2003 Server SP1 with dual
> CPU, looks like one of the process is blocked. In a lot of cases,
> the whole DB is blocked if this process is holding important locks.
>
> Looks like this issue was discussed in the following thread a few
> month ago, but didn't seem to have a solution mention. I would liek
> to know if there is a patch for this already?
>
> http://archives.postgresql.org/pgsql-performance/2006-
> 03/msg00129.php
>

There have been some fairly extensive changes in the semaphore code for
8.2. Any chance you can try the cvs snapshot version and see if the
problem exists there as well?

//Magnus

Hi Magnus,

Sure, I could try that out. Is there a place to download a 8.2 image to
bypass setting up the tool chain?

-Wei


On 9/4/06, Magnus Hagander <mha@sollentuna.net> wrote:
>
> > Hi,
> >
> > We are seeing hanging queries on Windows 2003 Server SP1 with dual
> > CPU, looks like one of the process is blocked. In a lot of cases,
> > the whole DB is blocked if this process is holding important locks.
> >
> > Looks like this issue was discussed in the following thread a few
> > month ago, but didn't seem to have a solution mention. I would liek
> > to know if there is a patch for this already?
> >
> > http://archives.postgresql.org/pgsql-performance/2006-
> > 03/msg00129.php
> >
>
> There have been some fairly extensive changes in the semaphore code for
> 8.2. Any chance you can try the cvs snapshot version and see if the
> problem exists there as well?
>
> //Magnus
>