[Linux-cluster] Postgres-cluster with shared storage.

[Mikko Partio]

On Mon, Jul 26, 2010 at 10:56 PM, Arun Kp <arunkp1987 at gmail.com> wrote:
>
> When the cluster try to relocate (Stopping in the Active node and starting
> in the Passive node) the Postgresql service  works fine if there is no
> client connection .But when one client is connected it will take 5 minutes
> for relocating.Here we have given 5 minutes as shutdown wait time in
> cluster.Some times relocating getting failed due to stopping of postgresql
> server taking too much time
> We have given the following options for postgresql in the cluster
>
> Name :DB-igvtdc
> Configdir:/Storage   where our data directory resides(In SAN).
> Postmaster options : -p 5433 -h 172.31.1.20 -m
> shutdown wait time :300 s
>

In your init script (/etc/init.d/postgresql or similar) check what is the
actual command that is used to shut down the database. My guess is that it's
something like

pg_ctl -D /path/to/database stop

Then just add switch "-m immediate" so that the database is not waiting for
all clients to disconnect before shutting done. More information on pg_ctl
syntax: http://www.postgresql.org/docs/8.4/static/app-pg-ctl.html

Regards

Mikko
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20100727/204ac09c/attachment.htm>