[asterisk-users] Asterisk sip_autodestruct messages - extensions locked

[Stefan Viljoen]

Hi guys

 

Does anybody have any opinion on what causes tens of thousands of these
messages per hour to pop up in the CLI:

 

[Jun 30 14:24:59] WARNING[2209]: chan_sip.c:4057 __sip_autodestruct:
Autodestruct on dialog '7e9597ae6ce95fef23374f4b380a9b70 at 192.168.0.1:5060'
with owner SIP/1148-0005bb2d in place (Method: BYE). Rescheduling
destruction for 10000 ms

[Jun 30 14:25:01] WARNING[2209]: chan_sip.c:4057 __sip_autodestruct:
Autodestruct on dialog '6faefcc24547f1e774864ca87e3ff335 at 192.168.0.1:5060'
with owner SIP/1028-0005bb3f in place (Method: BYE). Rescheduling
destruction for 10000 ms

[Jun 30 14:25:02] WARNING[2209]: chan_sip.c:4057 __sip_autodestruct:
Autodestruct on dialog '0d64480f4052a9e9054153552f1af7ba at 192.168.0.1:5060'
with owner SIP/1412-0005bb5d in place (Method: BYE). Rescheduling
destruction for 10000 ms

 

Symptoms are that an extension will dial, converse and hang up, but then be
unable to dial for up to four minutes after the initial call that extension
made - all the while thousands of the above messages scroll by in the CLI.
Along with then hundreds of "too many calls" warnings for each extension
which are limited to 1 call per extension at a time in sip.conf

 

I've googled intensively, no AGI is being run, hangup literally calls
Hangup(). I've already disabled and unloaded CDR-TDS, and ODBC CEL and CDR
logging to MySQL (in case db issues are causing a lock or something when
writing CDRs)

 

I see on big G that many many people have had this issue, but nowhere is
there any kind of concrete end result or answer to any of the questions.
This has started to happen at our site this Monday... it comes and goes
driven by no factor I can determine, except maybe workload. Server load
average when this is going on is 8 in top for a quad-Core i7 machine with
Centos 7 and 8GB of RAM, but the machine has hit 16 as load avg without this
happening. It also sometimes happens with the load average at 4 or less...

 

The particular server has been running undisturbed for about three years and
has handled tens of millions of calls, and only this Monday started
exhibiting this behaviour. The dialplan was last changed about three months
(and about 3 million calls) ago.

 

Any body got ANY advice or ideas where I can begin to diagnose this?

 

Thanks,

 

Stefan

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20170630/8af392fb/attachment.html>