In my experimenting with fallback routes, I have come across
the
following behaviour:
Setup: Asterisk PBX with 1 FXS and 1 FXO port, and 3 SIP
phones. It
connects to a network with a bridge interface (br0) which
bridges a
wifi interface and 4 normal ethernet intefaces. Connected to
the
ethernet are a router and a DNS server.
(What I believe are the) relevant configuration parts:
extensions.conf:
[outbound]
exten => _9XXXXXXXXXX,1,Dial(SIP/0030${EXTEN:1} sipprov1)
exten => _9XXXXXXXXXX,2,NoOp($)
exten => _9XXXXXXXXXX,3,Dial(Zap/2/${EXTEN:1})
sip.conf:
[general]
registerattempts=0
srvlookup=no
register => username:passwordsip.voipbuster.com
[sipprov1]
type=friend
username=username
host=sip.voipbuster.com
context=default
insecure=port,invite
secret=password
Case 1: No network connection (network interface exists
(br0), but no
ethernet cable is plugged in. Thus we have no DNS and no
actual access
to Internet.
FXS Try: Dialing from the FXS phone, I get a ~10 seconds
before the
attempt over SIP fails.
Case 2: I am connected to the network, and just have bad
username/password for VoIPBuster. It show shows as
registered in "sip
show registery"!
FXS Try: Dialing from the FXS phone, I get almost no time
before SIP
fails with "CONGESTION". (Error 500 from
voipbuster.com), and the call
goes on over Zap.
SIP Try: Calling from X-Lite and hanging-up before the
called number
answers results in immediate hangup.
Case 3: While the network was working, I remove the ethernet
cable.
Now the system has no connection, but doesn't know it at
application
level. We were also registered with VoIPBuster before the
cable was
removed.
FXS Try: Dialing from the FXS phone, I get a >100 seconds
delay before
the first dial fails (with CHANUNAVAIL). This is the same
even if the
"sipprov1" shows as "Unregistered" in
"sip show registry"
SIP Try: Dialing from a SIP phone, I get the same >100
seconds delay
before the first dial fails, but now I have an extra
problem. When the
call gets through to Zap, I press Hangup on the SIP phone
(X-Lite) and
Asterisk takes the "Hangup" packet after 30 - 50
seconds.
The above suggest to me that there is something fishy with
DNS
handling inside chan_sip, which forces all SIP handling to
wait. Am I
doing something wrong (besides pulling the cable)? Has
anyone else
observed this?
_______________________________________________
--Bandwidth and Colocation Provided by http://www.api-digital.c
om--
asterisk-dev mailing list
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-dev
|