[OpenSIPS-Users] Opensips loses location table contents

Schneur Rosenberg rosenberg11219 at gmail.com
Mon Aug 20 20:00:05 CEST 2012


It happened again just out of the nowhere, please can someone please
suggest what can cause it, I'm going crazy already.

On Sun, Aug 19, 2012 at 12:48 PM, Schneur Rosenberg
<rosenberg11219 at gmail.com> wrote:
> I'm having a interesting scenario, every once in a while my OpenSIPS
> location table begins losing its contents, I lose like 100 at a time,
> then it might regain some and then loses again another 100 or so.
>
> I have 2 servers one active and one standby via Heartbeat, heartbeat
> shuts and turns on opensips when becoming the active node,  the MYSQL
> databases are replicated as a MASTER/MASTER replication.
>
> First time this happened was when I tested a updated OpenSIPS script,
> so I turned on OpenSIPS on the standby node, I did not close OpenSIPS
> right away and thats why I suspected that the standby node kept on
> erasing lines once they expired and the active node kept on adding
> rows, this was a full fledged  OpenSIPS war :-)
>
> But this same thing happened last week again for no reason, I did not
> touch the standby server at all, all of the sudden we got calls from
> customers that their phones are acting weird, going online and offline
> all the time, sure enough I saw that the location table was down to
> maybe 20 registrations, I rebooted the master and it did not help, I
> rebooted the slave too and it did not help so I did a
> /etc/init.d/heartbeat stop and the slave fired up and everything went
> back to normal, within a few minutes all registrations were all back
> up.
>
> I then ran a apt-get upgrade and one of the applications that were
> updated was mysql server, so I said maybe that will fix the problem, I
> rebooted the master and heartbeat fired up OpenSIPS and everything was
> fine till today.
>
> Today I by mistake ran  opensipsctl fifo lb_list on the slave
> (standby) server and sure enough it gave me a error because opensips
> was not running, but 30 minutes later I got a call from a customer
> that his lines were down.
>
> Could it be that be merely running "opensipsctl fifo lb_list"
> something would of caused this issue? I did not run "ps aux | grep
> opensips" because I right away rebooted the servers, is there anything
> else that could cause such symptoms?
>
> thanks
> S. Rosenberg



More information about the Users mailing list