[Asterisk-Dev] CVS-HEAD: Weird "infinite ring" problem

[Brian Capouch]

Just in case this symptom list triggers someone's "Aha," I post this.

Symptom: incoming call triggers weird things, most notably stations that 
will start ringing and won't quit, even when the handset is lifted. 
This has been a problem for me since the jitterbuffer code was 
committed.  Reverting to CVS-HEAD from before that date (3/18?) fixes 
these problems up instantly.

Example 1: x86, Incoming analog call on X100P, triggers Dial to two Zap 
FXS ports, several IAX channels and a local SIP phone.  Randomly but at 
least once a day, a call will come in, and the Zap stations ONLY (none 
of the others) will start ringing just like normal.  But at that moment 
the system is hosed, the Zap channels never quit ringing until a CLI 
"stop now" is issued.  Long lag (~30 secs) after issuing stop now and 
the server actually going down.  Today, despite the fact that ringing 
phones trigger and intense psychological reaction in me, I did some 
playing around while those phones were doing their continuous-ring 
thing.  Nothing else looked amiss on the system, and I could even come 
on on other channels and do things.  Note that the other phones that are 
supposed to ring don't ever do so.  I can't leave this system in 
production because the caller never hears any ring when this phenom 
hits, and I don't want to miss calls.

Example 2: Just noticed with 04/07 version.  IAX call comes in to my 
Linksys WRT54GS server, which has dialplan entry to ring a pair of SIP 
phones.  One of them rings but the other doesn't.  I pick up the ringing 
phone and talk.  Approximately 30 seconds later, the *other* phone 
commences ringing, but nothing shows on the CLI and the "sip show peer" 
entry is as if the phone were there idle.  It will ring for about 30 
seconds and then stop.

The 30-second thing is weird.  The infinite ringing is weird.  I'm going 
to suspect something hosed in the SIP driver.  The x86 system runs just 
fine with old instance of the server and latest Zap drivers, so I don't 
think it lies in the CVS-HEAD version of Zap.

FYI, don't know enough to dig the code, the system isn't crashing, but I 
wanted to document this before someone else has this happen and spends 
hours thinking it's a config issue.

B.