[OpenSIPS-Users] opensips dispatcher looses media servers as unavailable after a while, when they alive

Anton VG anton.vazir at gmail.com
Wed Oct 9 13:19:48 CEST 2013


Hello!
Please help me resolving the nasty issue,

OpenSIPS have been setup as a proxy in front of 3xFreeSWITCH servers
passing managing all of the requests to them for further handling,
including
REGISTRATION requests, using DISPATHER

the problem is:
After a certain while (may vary from minutes to days) OPENSIPS treat the
remote media servers as DEAD

root at sipproxy:/# opensipsctl dispatcher dump
SET_NO:: 1
SET:: 1
        URI:: sip:192.168.100.10:5060 flag=P
        URI:: sip:192.168.100.11:5060 flag=P
        URI:: sip:192.168.100.12:5060 flag=P

and even unable to SIP-PING them via
opensipsctl ping sip:

root at sipproxy:/# opensipsctl ping sip:192.168.100.10:5060
408

at the same time servers are alive, and easily pingable if external tool is
used

root at sipproxy:~# sipsak -s sip:62.122.137.50:5060 -v
SIP/2.0 200 OK
... (cut)...

but after opensips is restarted (or sometimes by itself in a few minutes!)
- everything becomes fine... for a while again.

root at sipproxy:/# service opensips restart

root at sipproxy:/# opensipsctl dispatcher dump
SET_NO:: 1
SET:: 1
        URI:: sip:192.168.100.10:5060 flag=A
        URI:: sip:192.168.100.11:5060 flag=A
        URI:: sip:192.168.100.12:5060 flag=A

Tried opensips 1.10 and 1.9 with the same luck.

In the opensips logs there are no any unusual errors, no complaints to
memory exhaustion or connection limit. Nothing special.

opensips and media servers are running inside the VZ containers (if that is
important), also no notes any overlimits
OS Debian 7

Would be grateful for any clue.

Anton.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20131009/227bf77c/attachment.htm>


More information about the Users mailing list