DIALPLAN STUCK ON AMD does not trigger hangup until rtp timeout on sip trunk provider

We have answering machine detection (AMD_APP) running and on our dialer roughly 1 out of every 200 or so calls gets stuck on the AMD almost indefinitely and eventually return HANGUP after the sip trunk provider has an RTP timeout, our internal timeout does nothing here on triggering the hangup.

I set under sip settings RTP Timeout and RTP HOLD TImeout to 60 seconds.

We originate from a local context to a pjsip and everything works fine 99% of the time, but every so often it looks like the channel gets hung up on somewhere yet the dialplan still runs until it gets to the AMD command, then stays there until the trunk provider terminates.

Anything I can do to ensure if there are not two valid answered legs of the call and it gets to the AMD command that it terminates immidiately? I do not even know if that is the issue, all I can see is that the call stays open the whole time with our trunk provider and eventuals times out for an rtp timeout from their end not ours?

Any ideas what can be going on here? We are using an EC2 aws instance for our pbx, but like i said 99%+ of our calls go out no issue, it’s just very random. I verified our firewall ports were allowing ports 10000-20000 udp on both the local box and the aws firewall…

Any workarounds or something to try?

Initial Silence
2500
Greeting
1500
After Greeting Silence
800
Total Analysis Time
5000
Minimum Word Length
100
Maximum Word Length
5000
Between Words Silence
50
Maximum Number Of Words
3
Silence Threshold
256

Thanks!

This is the only output from log

– Channel PJSIP/voxtelesysshortpjsip-00000b20 left ‘simple_bridge’ basic-bridge <9a48bc49-0628-4550-9763-7e238df60997>
– Channel Local/[email protected];2 left ‘simple_bridge’ basic-bridge <9a48bc49-0628-4550-9763-7e238df60997>
== Spawn extension (outbound-dialer-custom, s, 5) exited non-zero on ‘Local/[email protected];2’
– AMD: Channel [Local/[email protected];1]. HANGUP

Asterisk version? There is an AMD issue with recent versions of 13 and 15:
https://issues.asterisk.org/jira/browse/ASTERISK-28143

On the Distro, you can downgrade Asterisk to 13.19.1 with:

yum -y downgrade asterisk13*-13.19.1
fwconsole restart

This seems like my issue. I am not using the distro, is there anyway I can just build the so and replace it? I am a novice anyway you can assist would be greatly appriciated!

I actually was able to just patch my asterisk 13.23.1 to use https://issues.asterisk.org/jira/secure/attachment/57968/amd.patch

Then just did a
make
make install

I am keeping my fingers crossed, but I think you nailed it. Thanks!

Will this patch be automatically included on future releases/versions of asterisk 13? I would like to ensure I don’t overwrite the patch if I upgrade.

Thanks!

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.