[Asterisk-Users] Asterisk crashes on special Transfer with MGCP/ATA 186

[brett-asterisk at worldcall.net]

Thomas Dingermann wrote:

> Hi all,
>
> i am using CVS-HEAD-08/13/04-12:00:00-BRI-stuffed-0.1.0-RC4a with 
> Cisco ATA-186 3.1.1 atamgcp
>
> We are used to make an special ;) blind transfer like 
> (Flash)Number(Hangup before anyone answers or ring).
> Then * crashes (see below) if the man in the middle is an 
> cisco-ata-186-mgcp
>
> If one waits until the last one rings, then hangup, everything is fine.
> If one waits until the last one  answers, then hangup, everything is 
> fine, too.
>
> Any hints?
>
>
> mgcp debug on:
>
>   -- Executing AGI("Zap/7-1", "nuller.agi") in new stack
>    -- Launched AGI Script 
> /home/kpj/pbx/var/lib/asterisk/agi-bin/nuller.agi
>    -- Accepting call from '01635571857' to '8551' on channel 0/1, span 3
>    -- AGI Script nuller.agi completed, returning 0
>    -- Executing Dial("Zap/7-1", "MGCP/aaln/1 at 192.168.1.26||") in new 
> stack
>    -- MGCP mgcp_request(aaln/1 at 192.168.1.26)
>    -- MGCP cw: 0, dnd: 0, so: 0, sno: 0
>    -- MGCP mgcp_new(MGCP/aaln/1 at 192.168.1.26-1) created in state: Down
>    -- Called aaln/1 at 192.168.1.26
>    -- MGCP/aaln/1 at 192.168.1.26-1 is ringing
>    -- Endpoint 'aaln/1 at 192.168.1.26-1' observed 'hd'
>    -- MGCP/aaln/1 at 192.168.1.26-1 answered Zap/7-1
> gw-bzo*CLI> mgcp debug on
> Usage: mgcp debug
>       Enables dumping of MGCP packets for debugging purposes
>    -- Endpoint 'aaln/1 at 192.168.1.26-1' observed 'hf'
>    -- Swapping 1 for 0 on aaln/1 at 192.168.1.26
>    -- MGCP Muting 1 on aaln/1 at 192.168.1.26
>    -- Started music on hold, class 'default', on Zap/7-1
>    -- MGCP mgcp_new(MGCP/aaln/1 at 192.168.1.26-0) created in state: Down
>    -- Endpoint 'aaln/1 at 192.168.1.26-0' observed '8'
>    -- Endpoint 'aaln/1 at 192.168.1.26-0' observed '7'
>    -- Endpoint 'aaln/1 at 192.168.1.26-0' observed 'hu'
>    -- Stopped music on hold on Zap/7-1
> Oct 15 13:32:58 NOTICE[100377]: chan_mgcp.c:1151 mgcp_fixup: 
> mgcp_fixup(Zap/7-1, Zap/7-1<MASQ>)
>    -- Swapping 0 for 1 on aaln/1 at 192.168.1.26
> Oct 15 13:32:58 WARNING[14350]: chan_mgcp.c:3033 handle_request: 
> Transfer attempt 
> failed                                                                  
> ______

Hey There,
I had a similar problem running CVS HEAD 02/09/05. An Attended transfer 
that wasn't completed caused the channel to lockup. with a bunch of:
Mar 16 14:12:12 WARNING[8904]: channel.c:523 ast_channel_walk_locked: 
Avoided deadlock for 'MGCP/aaln/1 at 192.168.10.5-1', 10 retries!

messages.. Did you ever find a solution??

Thanks,
Brett