Asterisk 11.11.0 on CentOS 6.5, installed from RPMs. OpenSIPS fronting
Asterisk from a Tekelec T9000.
I'm accumulating stuck channels.
I'm googling now and I recognize that Friday afternoons are the worst time
to ask questions, but I'm getting desperate because this is keeping me
from rolling a system out to production. (Yup, I know. Who rolls out a
system on a Friday afternoon...)
They all show (core show channels) that they are running one of my AGIs,
so I suspect something changed in processing AGIs from 1.2(!). (I've
recently changed from using 'signal()' to using 'sigaction()' if
that
rings any bells with anyone.)
That's puzzling enough, but killing the AGI process from the shell
doesn't
help.
Neither does 'channel request hangup SIP/a.b.c.d-00000031'
Neither does 'channel redirect SIP/a.b.c.d-00000031 default,s,3' where
'default,s,3' is a hangup().
Neither of these commands log any errors.
The only other clue may be this message:
Autodestruct on dialog '1405009120-21730426 at T9000' with owner
SIP/a.d.b.c-00000031 in place (Method: BYE). Rescheduling destruction for
10000 ms
for each of the hung channels every 30 seconds or so.
I haven't identified what callers are doing to reproduce the error
reliably yet.
Any clues or suggestions?
--
Thanks in advance,
-------------------------------------------------------------------------
Steve Edwards sedwards at sedwards.com Voice: +1-760-468-3867 PST
Newline Fax: +1-760-731-3000