IAX and AMD appears to leave dead channels open.

[AIRAM]

We started having issues on one of our test servers were we started getting more users in and then quality issues started to occur and when dialing out were getting a recording saying that "the call could not be placed as dialed".

After further troubleshooting and getting more channels from our provider we noticed that actually the channels were left open for hours. Lowering the auto dial level (obviously uses less channels) does seem to help but what concerns me is why are the channels stuck the last application on them is AMD. Here's the channel info:

asterisk-vicidial*CLI> show channel IAX2/binfone00-6
-- General --ial*CLI>
Name: IAX2/binfone00-6
Type: IAX2
UniqueID: 1151517023.2136
Caller ID: XXXXXXXXXX
Caller ID Name: V0628115022000029442
DNID Digits: (N/A)
State: Up (6)
Rings: 0
NativeFormat: 4
WriteFormat: 64
ReadFormat: 64
1st File Descriptor: -1
Frames in: 102
Frames out: 0
Time to Hangup: 0
[b] [color=red]Elapsed Time: 1h16m2s[/color][/b]
Direct Bridge: <none>
Indirect Bridge: <none>
-- PBX --
Context: default
Extension: 8369
Priority: 2
Call Group: 0
Pickup Group: 0
[color=red] [b]Application: AMD[/b][/color]
Data: 2000|2000|400|5000|120|25|5|256
[color=red][b]Blocking in: ast_waitfor_nandfds[/b][/color]
Variables:
BRIDGEPEER=Local/XXXXXXXXXX@default-55d4,2
DIALEDPEERNUMBER=MyAccount:MyPass@iax.binfone.com/1XXXXXXXXXX

CDR Variables:l*CLI>
level 1: clid="V0628115022000029442" <XXXXXXXXXX>
level 1: src=XXXXXXXXXX
level 1: dst=8369
level 1: dcontext=default
level 1: channel=Local/XXXXXXXXXX@default-55d4,1
[color=red][b]level 1: lastapp=AMD[/b][/color]
level 1: lastdata=2000|2000|400|5000|120|25|5|256
level 1: start=2006-06-28 11:50:25
level 1: answer=2006-06-28 11:50:25
level 1: end=2006-06-28 11:50:25
level 1: duration=0
level 1: billsec=0
level 1: disposition=ANSWERED
level 1: amaflags=DOCUMENTATION
level 1: uniqueid=1151517023.2136


It appears as if it was ended but the channel just stays stuck there. Seems like after a while they are killed but it seems to be affecting the quality of our calls.

Any ideas what might be going wrong?

[mflorell]

I've never seen that before, although we have only run AMD for a few small campaigns. What codec are you using for those calls? can you see anything in the message file with debug turned on for those calls?

[AIRAM]

We are using Ulaw but we have tried GSM as well. We are not really seeing any output since we are only getting it from "concluded" AMD transactions.

This ones some how seem to be just waiting for some type of input or something. I'm specialy concerned about the line that says Blocking in: ast_waitfor_nandfds. Since is present on all those "zombie" channels.

Anyway we are enabling verbose for AMD and testing further to check if we can find anything else.

[BonTheOne]

I know this is an old thread, but I was wondering if anyone else has seen this issue or found a solution for it, as I seem to be experiencing the same problem?

Thanks,

Ian