Hi
We have OCS R2 setup to use Asterisk to make calls out to the PSTN (via the Avaya phone system). This has been working great for a few years. All of a sudden this weekend we started seeing the error below. Any ideas what is causing this? The call starts when an OCS user dials a number, the call is passed off to Asterisk who then should pas the call to the Avaya PBX as H.323. Worked fine up to now.
(FYI - I removed the real IP address and replaced it with x’s below)
== Everyone is busy/congested at this time (1:0/0/1)
– Auto fallthrough, channel ‘SIP/xxx.xxx.xxx.xxx-09606970’ status is ‘CHANUNAVAIL’
You are trying to make the call as SIP, not H.323. You possibly don’t have chan_sip loaded.
If you really meant SIP, it probably doesn’t like the IP address. Is it routable from the machine? Is it actually well formed?
I mean, an OCS users dials a number, it gets sent to Asterisk as SIP and then asterisk converts to H323 and sends to the Avaya PBX.
When I do a SIP show peers I see the OCS mediation server and the status is OK, using port 5060. Any tips on troubleshooting would be greatly appreciated. 
I can telnet to the OCS server’s 5060 port from Asterisk too.
Type Description Devicestate Indications Transfer
USTM UNISTIM Channel Driver no yes no
SIP Session Initiation Protocol (SIP) yes yes yes
Local Local Proxy Channel Driver yes yes no
Phone Standard Linux Telephony API Driver no yes no
Agent Call Agent Proxy Channel yes yes no
H323 The NuFone Network’s Open H.323 Channel no yes no
----------1*CLI>
6 channel drivers registered.
My mistake. You didn’t show the channel it was trying to dial, so I picked on the only channel that was actually included, but that was the incoming channel.
You may need to do “core set verbose 3”, but one way or another, you need find the channel name that is actually being dialed. That channel name will reference something that doesn’t exist.
Quite often, the primary error will be reported before the channel unavailable condition.
One other thing. If you are seeing IP addresses in incoming channel names, it means you have allowguest set to yes and you don’t have a valid match in sip.conf. You don’t want that situation if the machine is reachable from the internet.
Here’s a call sample but let me know if you need more to see. I edited out the real server names and domain names and phone numbers and replaced with something that won’t ID the company, etc. But you should get the picture
== Using SIP RTP CoS mark 5
– Executing [+17145551212@default:1] AGI(“SIP/Ocs-mediation-server-09663768”, “2ldap.agi,tel:+122436,+17145551212”) in new stack
– Launched AGI Script /var/lib/asterisk/agi-bin/2ldap.agi
– <SIP/Ocs-mediation-server-09663768>AGI Script 2ldap.agi completed, returning 0
– Executing [25417@core:1] Set(“SIP/Ocs-mediation-server-09663768”, “CALLERID(number)=2487122436”) in new stack
– Executing [25417@core:2] Dial(“SIP/Ocs-mediation-server-09663768”, “H323/25417@Avaya,120,tr”) in new stack
– Requested transfer capability: 0x00 - SPEECH
– Called 25417@Avaya
== Everyone is busy/congested at this time (1:0/0/1)
– Auto fallthrough, channel ‘SIP/Ocs-mediation-server-09663768’ status is ‘CHANUNAVAIL’
If allowguest is missing from the sip.conf file, does that mean that it’s basically a ‘yes’ anyway if that’s the default setting for allowguest? I don’t see allowguest in the sip.conf file.
Thanks for your replies…it’s working now. Crazy…the card in the PBX was freaking out. We reset it and now the calls can pass through correctly. I didn’t check that earlier because the calls didn’t seem to even try to get to the pbx. Seemed like it was failing at the OCS side.
allowguest did and probably still does default to true. The argument is that it makes it easier for people to get an evaluation configuration working, although that applies more exactly to the example files.