[OpenSIPS-Users] opensips Crashing v2.2.2
bogdan at opensips.org
Tue Jan 17 05:10:52 EST 2017
Sorry for the late reply. The backtraces show a really strange situation
- some processing context getting to NULL somewhere along the handling
of that 503 reply.
What I suspect is that you are using in failure route some functions
that are reseting this processing context - could you (off list) send me
the failure route which is used for this 503 reply ?
Also, when you say the carrier is responding with a shorter message -
what you mean ? shorter as length ? or ?
OpenSIPS Founder and Developer
On 01/16/2017 02:57 PM, Richard Robson wrote:
> I've done some more testing and the carrier is now responding with a
> shorter message, despite them saying nothing was wrong. and the
> crashing isn't happening.
> Therefore I'm unable to reproduce the original error on our live system.
> I've put the node back into service and I'm monitoring the situation.
> On 06/01/2017 12:19, Richard Robson wrote:
>> I've some more back traces.
>> this could be a resource problem. The back story is we want to check
>> for channel limits and Rate limits for all calls. Channel limit hits
>> should failover to the next gateway and rate limits should delat for
>> up to 2 seconds before failing over to the next gateway. This is all
>> working ok as far as I can tell. We have one provider that has 600
>> channels and cps of 25. We now have a second provider with 3 nodes of
>> 100 channels and 4 cps. All nodes are being checked for Channels and
>> CPS. I noticed that the failover of GW was not being checked and
>> added new route to do the checking. Also to confuse things more it
>> looks like one of the new carriers nodes may be incorrectly
>> configured asnd we are getting failures (503 Service unavailable).
>> Yesterday we were using all four gateways and no crashes, but after
>> adding in the new route for the failover checking it started to crash
>> this morning. the logs indicate that the failover is working correctly.
>> We are checking the outbound calls twice once for the customer limits
>> and then for the carrier limits. This has been fine for the last few
>> days, since the system was updated ( it was only carrier limits
>> checked before)
>> i've removed the suspect node and the extra route, but I notice the
>> error from the node in the backtraces with an address out of bounds
>> I've more back traces if you require
>> On 06/01/2017 09:30, Richard Robson wrote:
>>> we are getting a crash in opensips. V2.2.2
>>> bt all:
More information about the Users