[asterisk-dev] 900 second call drop - 420 Bad Extension error w/ SIP, IOS GW

[Jared Mauch]

Greetings,

I've been trying to track down an odd issue I've seen in versions
after 1.6.1.4 (including 1.6.2).

We have an IOS gateway that connects us to the PSTN for performing SIP
inbound/outbound termination.

What we see is when a call originates on the PSTN (so comes in the PRI
and is handed to Asterisk) after 900 seconds the call is dropped.

I'm not sure what is triggering this exactly. I've looked at SIP traces and
getting the exact nuance sorted out is tough, but what I see is something
similar to this:

    -- Got SIP response 420 "Bad Extension" back from x.x.2.115
  == Spawn extension (verio-int, 600, 4) exited non-zero on 'SIP/dns.name.hidden-00000064'


The actual message received is usually something like this:

SIP/2.0 420 Bad Extension
Via: SIP/2.0/UDP x.x.27.16:5060;branch=z9hG4bK21db1330;rport
From: <sip:2149151350 at x.x.27.16>;tag=as22d8cfaf
To: <sip:2149151356 at x.x.2.115>;tag=6E691D88-4A
Call-ID: C8B6CF60-D86311DE-93A08FDB-20B8ACA9 at x.x.2.115
CSeq: 102 INVITE
Unsupported: timer
Content-Length: 0


I think I have a full call trace available in pcap format for those that
may be interested in looking at this.  I'm wondering what would cause the
drop at exactly 900 seconds, as that seems to be some timeout/timer, and
if you had any tips of where I should be looking at in the code to identify
this issue.

The IOS gateway has been up for a few years prior to this issue without
any problems.  I did upgrade the code to something more recent to attempt
to determine if it was an IOS defect but it does not seem to be.

Thanks for any help.

	- Jared

-- 
Jared Mauch  | pgp key available via finger from jared at puck.nether.net
clue++;      | http://puck.nether.net/~jared/  My statements are only mine.