[Asterisk-Users] Program loop & segfault in heavily used Unicall channels

[Guillermo Freige]

Hi:

This is a slightly modified repost as I never saw the original post in the 
list.

I'm having problems with Unicall channels in a heavily used asterisk. The
box is using 2 E1 lines in mfcr2, with both incoming and outgoing calls,
uning Argentina R2 variant. Besides the already known problem when the
outgoing calls failing and blocking a channel, now I have loop problems.
When I reach around 40 channels in use, and after some time, some channels
generate an endless loop of error messages, CPU goes to 99%,  locking the
channel. Also I got segfaults when Asterisk is loaded, with or without 
channels locked, sometimes after 2 hours, usually every hour, and sometimes 
multiple times after a segfault (4 or 5 times in less than 10 minutes) until 
it stabilizes. I never saw this problem during testing, not like the 
previously mentioned ougoing call problem, that happen despite the level of 
usage.

First it generates a short bunch of this message:
Dec  9 14:40:18 WARNING[25526315]: We're UniCall/17-1, not UniCall/11-1
Dec  9 14:40:18 WARNING[25526315]: ast = 0x8283328, p->owner = 0x81bd7b8

Lately it generates the following messages in an endless loop:
Dec  9 15:02:35 WARNING[4046889]: We're UniCall/17-1, not
Dec  9 15:02:35 WARNING[4046889]: ast = 0x8243720, p->owner = 0x8256478

Now I've changed this message from WARNING to DEBUG to not to clog the
logfile, but the problem remains the same.

Any ideas??

Guillermo

_________________________________________________________________
MSN Amor: busca tu ½ naranja http://latam.msn.com/amor/