Hi all, I have VOIP network built with MGCP endpoints.The manufacturer of endpoints is ASKEY. I downloaded latest Asterisk software and found it very useful for me. I configured it and it seems taht everything works OK when I am testing it with one or two endpoints. After that I tried to move Asterisk to working network and replace existing call manager. It starts working and calls are proceeding but after a while I could not hear a dialtone and saw in logs the following: Nov 27 11:40:57 WARNING[10251]: File chan_mgcp.c, Line 2127 (handle_hd_hf): Unable to create switch thread: Interrupted system call I looked in chan_mgcp.c file and saw that this error occures after pthread_create functions and it means that this "system call was interrupted permaturely with a signal before it was able to complete". Please, help me to resolve this problem. Best regards, Sergi Gabunia -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20031127/e7bd2c97/attachment.htm
Hi, I forgot to say that I have about 300 MGCP endpoints in my real network. Best regards, Sergi Gabunia ----- Original Message ----- From: Sergi Gabunia To: asterisk-users@lists.digium.com Sent: Thursday, November 27, 2003 12:05 PM Subject: [Asterisk-Users] MGCP problem Hi all, I have VOIP network built with MGCP endpoints.The manufacturer of endpoints is ASKEY. I downloaded latest Asterisk software and found it very useful for me. I configured it and it seems taht everything works OK when I am testing it with one or two endpoints. After that I tried to move Asterisk to working network and replace existing call manager. It starts working and calls are proceeding but after a while I could not hear a dialtone and saw in logs the following: Nov 27 11:40:57 WARNING[10251]: File chan_mgcp.c, Line 2127 (handle_hd_hf): Unable to create switch thread: Interrupted system call I looked in chan_mgcp.c file and saw that this error occures after pthread_create functions and it means that this "system call was interrupted permaturely with a signal before it was able to complete". Please, help me to resolve this problem. Best regards, Sergi Gabunia -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20031127/d944a0a5/attachment.htm
> Message: 1 > From: "Sergi Gabunia" <sg@wanex.net> > To: <asterisk-users@lists.digium.com> > Date: Thu, 27 Nov 2003 12:05:15 +0400 > Subject: [Asterisk-Users] MGCP problem > Reply-To: asterisk-users@lists.digium.com > > I have VOIP network built with MGCP endpoints.The manufacturer of > endpoints is ASKEY. I downloaded latest Asterisk software and found it > very useful for me. I configured it and it seems taht everything works > OK when I am testing it with one or two endpoints. After that I tried to > move Asterisk to working network and replace existing call manager. It > starts working and calls are proceeding but after a while I could not > hear a dialtone and saw in logs the following: > Nov 27 11:40:57 WARNING[10251]: File chan_mgcp.c, Line 2127 > (handle_hd_hf): Unable to create switch thread: Interrupted system call > > I looked in chan_mgcp.c file and saw that this error occures after > pthread_create functions and it means that this "system call was > interrupted permaturely with a signal before it was able to complete".=20 > > Please, help me to resolve this problem. > > Best regards, > Sergi Gabunia >I am using v1.29 of chan_mgcp.c with the askey unit and I can see a similar problem with memory not being released after off hook/on hook transition. Anybody fill me in on what debugging data would be useful in identifying this problem? darren
The latest cvs mgcp code seems to be slightly broken. I have two gateways
and both of them only report their ip address to * instead of a hostname.
When I audit the endpoints, * properly places the ip address in brackets and
the audit is successful. The phones connected to the gateway get dial tone
when I take them off hook, and I can originate calls from those phones. If
I attempt to originate a call from my sip phones to these same mgcp
endpoints, * does not place the ip address in brackets, and of course the
call fails. This was working before, but now it is broken.
-- Executing Dial("SIP/9999-0ccc",
"MGCP/aaln2/@66.17.13.240") in new stack
Apr 23 13:23:40 NOTICE[376847]: chan_mgcp.c:1457 find_subchannel: Gateway
'66.17.13.240' (and thus its endpoint 'aaln2/') does not exist
Apr 23 13:23:40 WARNING[376847]: chan_mgcp.c:3205 mgcp_request: Unable to
find MGCP endpoint 'aaln2/@66.17.13.240'
Apr 23 13:23:40 NOTICE[376847]: app_dial.c:554 dial_exec: Unable to create
channel of type 'MGCP'
Thanks,
Brad White
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.digium.com/pipermail/asterisk-users/attachments/20040423/264d6cb4/attachment.htm
I seem to have fallen off the list, but saw a reply in the list archives:>Please add more info. Use 'ngrep port 2727' in order to see the trafficbetween the MG and MGC and post that results >here. -- Juanjo sin .sig This is a misunderstanding of the problem. There is not any traffic between the MG and MGC to look at because asterisk can't determine the endpoint, re:> > -- Executing Dial("SIP/9999-0ccc", "MGCP/aaln2/@66.17.13.240") in new >stack > Apr 23 13:23:40 NOTICE[376847]: chan_mgcp.c:1457 find_subchannel: > Gateway '66.17.13.240' (and thus its endpoint 'aaln2/') does not exist > > Apr 23 13:23:40 WARNING[376847]: chan_mgcp.c:3205 mgcp_request: Unable > to find MGCP endpoint 'aaln2/@66.17.13.240' > Apr 23 13:23:40 NOTICE[376847]: app_dial.c:554 dial_exec: Unable to > create channel of type 'MGCP' This is a sip client trying to call a mgcp endpoint. The very same mgcp endpoint gets sent dialtone from * and can also originate calls to either sip clients, or to * itself(the "500" demo for example). I'm not a programmer, but the problem as I see it is that * does place brackets around the ip address when the mgcp endpoint originates a call and when * audits the endpoint, but * does not place brackets around the ip address when the mgcp endpoint is the destination of a call. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20040505/bb8ff437/attachment.htm
Turns out this was a typo in my extensions.conf file all along. Many thanks to the person who pointed it out. The answer was staring me in the face the entire time, but I just couldn't see it. Apologies to all.... -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20040507/1ffe4b9d/attachment.htm
Hello, I have noticed a little problem in chan_mgcp.so. After a few unsuccessful attempts to call an endpoint using MGCP/aaln/1@10.20.0.2 I have noticed the following on the system running asterisk using netstat. udp 50524 0 XX.XX.85.XX:2427 0.0.0.0:* 2666/asterisk This line shows that there are 50524 bytes waiting in the Recvq of the udp socket of asterisk. And the transactions are all timing out. Snip Retransmitting #4 transaction 63 on [10.20.0.2] Retransmitting #4 transaction 64 on [10.20.0.2] Retransmitting #5 transaction 63 on [10.20.0.2] Retransmitting #5 transaction 64 on [10.20.0.2] Aug 24 14:49:14 WARNING[24601]: chan_mgcp.c:595 retrans_pkt: Maximum retries exceeded for transaction 63 on [10.20.0.2] Aug 24 14:49:14 WARNING[24601]: chan_mgcp.c:595 retrans_pkt: Maximum retries exceeded for transaction 64 on [10.20.0.2] Aug 24 14:49:14 NOTICE[24601]: chan_mgcp.c:2261 handle_response: Transaction 64 timed out Aug 24 14:49:14 NOTICE[24601]: chan_mgcp.c:2283 handle_response: Terminating on result 406 from aaln/1@10.20.0.2--1 End snip. The system is GNU/Linux Debian 3.0r2 kernel 2.4.26. Asterisk CVS-HEAD-06/22/04-13:37:51 built by root@host.domain.xx on a i686 running Linux Any ideeas ??? Kiss Karoly
Any suggestions? The call does not pass through.
-- Making call to 1995@10.1.105.3.
== New H.323 Connection created.
-- 6129 is calling host 1995@10.1.105.3
-- Call token is ip$localhost/24274
-- Call reference is 24274
=*= In CreateRealTimeLogicalChannel for call 24274
-- externalIpAddress: 10.23.17.5
-- externalPort: 10238
-- SessionID: 1
-- Direction: IsReceiver
=*= In CreateRealTimeLogicalChannel for call 24274
-- externalIpAddress: 10.23.17.5
-- externalPort: 10238
-- SessionID: 1
-- Direction: IsTransmitter
=*= In CreateRealTimeLogicalChannel for call 24274
-- externalIpAddress: 10.23.17.5
-- externalPort: 10238
-- SessionID: 1
-- Direction: IsReceiver
=*= In CreateRealTimeLogicalChannel for call 24274
-- externalIpAddress: 10.23.17.5
-- externalPort: 10238
-- SessionID: 1
-- Direction: IsTransmitter
-- Sending SETUP message
0:14.798 H225 Caller:81830f0 h323pdu.cxx(1176) H225
No Q931 User-User Information Element,
Raw PDU:
08 02 de d2 0d .....
Q.931 PDU:
{
protocolDiscriminator = 8
callReference = 24274
from = destination
messageType = SetupAck
}
-- Ringing phone for "10.1.105.3"
0:16.080 H225 Caller:81830f0 h323pdu.cxx(1176) H225
No Q931 User-User Information Element,
Raw PDU:
08 02 de d2 6e 27 01 f1 4c 05 80 31 39 39 35 ....n'..L..1995
Q.931 PDU:
{
protocolDiscriminator = 8
callReference = 24274
from = destination
messageType = <110>
IE: 0x27 (39) = {
f1 .
}
IE: Connected-Number = {
80 31 39 39 35 .1995
}
}
0:18.376 H245:818cf28 h323.cxx(4082) H245
Received early start OLC, aborting fast start
=*= In CreateRealTimeLogicalChannel for call 24274
-- externalIpAddress: 10.23.17.5
-- externalPort: 10238
-- SessionID: 1
-- Direction: IsReceiver
-- Started logical channel: receiving G.729{sw}
-- channelsOpen = 1
-- Received RELEASE COMPLETE message...
-- Sending RELEASE COMPLETE
channelsOpen = 0
0:25.276 H245:818cf28 h323.cxx(3195) H245
Read error: Interrupted system call
0:26.398 H323 Cleaner h323.cxx(1542) H323
Connection ip$localhost/24274 terminated.
-- Congested link to 10.1.105.3
== H.323 Connection deleted.