[OpenSIPS-Users] opensips long shutdown time

[Dave Singer]

Thanks for the reply.

I would also lean to the DB as the problem but I have also experienced it
with about the same frequency when not using any database modules and low or
now calls.
So if you could point me to the patch you refereed to I would appreciate it.
I'll then run it through a scripted process of restarting to try to
reproduce and catch the problem (as time permits).

Thanks again for all your work with development and support!!

Dave (cando)

On Fri, Aug 27, 2010 at 2:02 AM, Bogdan-Andrei Iancu <bogdan at voice-system.ro
> wrote:

> Hi Dave,
>
> At sthutdown, opensips modules are mainly doing memory cleanup (dialog,
> transactions, caches, etc) or data flush to DB (dialogs, usrloc, etc).
>
> As memory cleanup is predictable as time and not resource consuming, I
> would say it is something related to DB flush - some DB ops that take
> random number of (mili) seconds, depending on the DB server load.
>
> A simple patch to report the shutdown duration of each module can be
> made if you consider it helpful - but I suspect the DB ops.
>
> Regards,
> Bogdan
>
> Dave Singer wrote:
> > Sometimes when I'm do a restart on opensips (using init.d script, with
> > some customization to handle this problem) opensips takes quite a bit
> > of time, like 30 - 50 seconds to stop. Other times it is very quick.
> > I'm using 1.6.2 and 1.6.3.
> > The servers are fairly busy, less then 200 calls per sec, with around
> > 1000 sustained open calls.
> > I am using dialog module and db mode is 3 (on shutdown) with
> > postgresql 8.4. However since sometimes it is very quick with the same
> > call volume it doesn't "feel" like this is the problem. Further I have
> > had it be slow to stop when testing a config that is not using dialog
> > at all and there are no active calls or any sip activity when I try to
> > stop it.
> > My hunch is slightly toward it is waiting on a transaction but not so
> > sure.
> >
> > The pertinent part of my init.d script customizations are for after
> > sending opensips the kill <pid> signal to every 0.1 seconds check if
> > it has stopped yet then start it back up.
> >
> > Needless to say, having it take 50 seconds to restart, and not
> > responding to sip traffic does not make customers happy when your
> > dealing with thousands of call. :(
> >
> > Any thoughts or things to try would be appreciated.
> >
> > Thanks
> >
> > cando
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Users mailing list
> > Users at lists.opensips.org
> > http://lists.opensips.org/cgi-bin/mailman/listinfo/users
> >
>
>
> --
> Bogdan-Andrei Iancu
> OpenSIPS Bootcamp
> 20 - 24 September 2010, Frankfurt, Germany
> www.voice-system.ro
>
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.opensips.org/pipermail/users/attachments/20100827/06ca8169/attachment.htm