[asterisk-users] Re: SIP/IAX peers UNREACHABLE and audio loss

[Edoardo Serra]

Hi all,
	I made some tests under heavy network load generated artificially 
moving files form server to server

I noticed a 3% packet loss in ping -f response form server involved in 
big data transfer (1 GB files through http)

I changed the network switch with a Cisco Catalyst 2950 and the packet 
loss with pings disapperead but the problem with REACHABLE / UNREACHABLE 
peers remains...

I did one more simple test
While Asterisk is stating the peer is UNREACHABLE I can ping (even -f) 
it without problem and without packet loss.

Could it be a problem in Asterisk ?

I'm using 1.2.13 on a gentoo
Kernel 2.6.20

Tnx again for help

Edoardo


Edoardo Serra ha scritto:
> Hi all,
>     I'm having a problem with some Asterisk servers interconnected with 
> each other using IAX (I also tried with SIP without solving the problem)
> 
> Sometimes, with apparently no reason, some peers become UNREACHABLE
> (I have qualify=yes in iax.conf) and REACHABLE again as soon as
> another qualify test is made.
> 
> Our users are also complaining about audio loss during their calls,
> apparently randomly, everything goes ok for days and bad for another few 
> days.
> 
> I strongly believe the 2 problems are strictly related because in the 
> logs I see REACHABLE / UNREACHABLE messages only for certains days
> without regularity.
> The days in wich i see a lot of messages are exactly the days with
> most of complaint about audio loss
> 
> I just noticed that timestamps of the logs (REACHABLE / UNREACHABLE)
> are quite always during business hours, this makes me think at somewhat 
> related to load (cpu load, badwidth load, calls load, etc...)
> 
> But, looking at hardware specs of our lan, servers and average load I 
> don't think they are over-stressed.
> 
> Our servers are all:
> 2 x Intel(R) Xeon(TM) CPU 3.20GHz
> 1 GB RAM
> 2 x IDE HDDs Software RAID 1
> Asterisk 1.2.13 with res_perl
> Gentoo Linux
> Some of them has a Sangoma card connected with an E1
> 
> Most ot these are on the same LAN, interconnected with a 1 GB switch
> (I don't think it should be a bandwidth problem).
> 
> Load averages of these server is varying from 0.5 to 1.0
> (I guess it should be ok)
> 
> On each server we don't have more than 50 concurrent calls
> (bridged SIP <-> IAX2 or IAX2 <-> ZAP)
> 
> Used codec is mostly G729
> 
> Sometimes on asterisk cli i see some messages like
> "Avoided initial deadlock for '0x9fd130', 10 retries!"
> I don't know if it could be somehow related.
> 
> Someone of you can point me in the right direction ?
> 
> Tnx in advance
> 
> Regards
> 
> Ing. Edoardo Serra
> WeBRainstorm S.r.l.
> 
> _______________________________________________
> --Bandwidth and Colocation provided by Easynews.com --
> 
> asterisk-users mailing list
> To UNSUBSCRIBE or update options visit:
>   http://lists.digium.com/mailman/listinfo/asterisk-users
>