Thread: restart error

[postgresql@stell ~]$ ps axuw | grep postmaster
502      11134  0.0  0.2 19292 3020 pts/5    S    Sep22   0:00 /usr/local/pgsql/bin/postmaster -D /var/pgsql/data
502      21676  0.0  0.0  4788  656 pts/12   S+   10:42   0:00 grep postmaster
[postgresql@stell ~]$ pg_ctl restart  -D /var/pgsql/data
waiting for postmaster to shut down............................................................... failed
pg_ctl: postmaster does not shut down
[postgresql@stell ~]$

yet postmaster is gone:
[postgresql@stell ~]$ ps axuw | grep postmaster
502      21717  0.0  0.0  4640  644 pts/12   R+   10:52   0:00 grep postmaster

server.log not updated...?
[postgresql@stell ~]$ ls -l /home/postgresql/server.log
-rw-rw-r--  1 postgresql postgresql 2343 Aug 23 09:48 /home/postgresql/server.log

How can I figure out what the problem is?

Ray Stell <stellr@cns.vt.edu> writes:
> [postgresql@stell ~]$ pg_ctl restart  -D /var/pgsql/data
> waiting for postmaster to shut down............................................................... failed
> pg_ctl: postmaster does not shut down
> [postgresql@stell ~]$

> yet postmaster is gone:

Looking at the code, it seems that pg_ctl is waiting for the postmaster
lock file (/var/pgsql/data/postmaster.pid) to disappear.  Perhaps you've
got some kind of file permissions bollix that prevented the postmaster
from removing that file?

            regards, tom lane

On Wed, Oct 04, 2006 at 01:28:37PM -0400, Tom Lane wrote:
> Ray Stell <stellr@cns.vt.edu> writes:
> > [postgresql@stell ~]$ pg_ctl restart  -D /var/pgsql/data
> > waiting for postmaster to shut down............................................................... failed
> > pg_ctl: postmaster does not shut down
> > [postgresql@stell ~]$
>
> > yet postmaster is gone:
>
> Looking at the code, it seems that pg_ctl is waiting for the postmaster
> lock file (/var/pgsql/data/postmaster.pid) to disappear.  Perhaps you've
> got some kind of file permissions bollix that prevented the postmaster
> from removing that file?


Turns out I had an open session in another shell I'd
forgotten about.  I tried again with similar results and
as I sent \q on the outstanding session, pm went down instantly:

[postgresql@wireworm global]$ pg_ctl stop -D /usr/pgsql/data
waiting for postmaster to shut down.................................................. done
postmaster stopped

Also, kill -15 on active psql sessions let postmaster go down quickly.

On Thursday 05 October 2006 16:26, Ray Stell wrote:
> On Wed, Oct 04, 2006 at 01:28:37PM -0400, Tom Lane wrote:
> > Ray Stell <stellr@cns.vt.edu> writes:
> > > [postgresql@stell ~]$ pg_ctl restart  -D /var/pgsql/data
> > > waiting for postmaster to shut
> > > down...............................................................
> > > failed pg_ctl: postmaster does not shut down
> > > [postgresql@stell ~]$
> > >
> > > yet postmaster is gone:
> >
> > Looking at the code, it seems that pg_ctl is waiting for the postmaster
> > lock file (/var/pgsql/data/postmaster.pid) to disappear.  Perhaps you've
> > got some kind of file permissions bollix that prevented the postmaster
> > from removing that file?
>
> Turns out I had an open session in another shell I'd
> forgotten about.  I tried again with similar results and
> as I sent \q on the outstanding session, pm went down instantly:

select * from pg_stat_activity;

is your friend.

Sean

OR you can use

pg_ctl -D <data-folder> -mf

Thanks,
----
Shoaib Mir
EnterpriseDB (www.enterprisedb.com)

On Fri, Oct 06, 2006 at 01:35:52AM +0500, Shoaib Mir wrote:
> OR you can use
>
> pg_ctl -D <data-folder> -mf

pg_ctl stop -D <data-folder> -mf

nice.  Had not found that in google, lots of kill -15 refs

thanks.

This is actually a fast mode of stopping the database server. You saw a lot of -15 as all the clients were forcefully disconnected as result of the -mf switch.

You can find details at --> http://www.postgresql.org/docs/8.1/interactive/app-pg-ctl.html

""Fast" mode does not wait for clients to disconnect. All active transactions are rolled back and clients are forcibly disconnected, then the server is shut down"

Thanks,
-----
Shoaib Mir
EnterpriseDB (www.enterprisedb.com)