[cisco-voip] Another busy signaling issue

[Andre Beck]

Hi,

after upgrading the CCM to the latest 4.1.3 SR4b (and migration to
another machine), users started complaining that incoming calls from
the PSTN don't singnal busy when the DID destination actually was busy.
The caller hears silence for approx. 15s and then gets an error tone
(which sounds like a busy, but actually is triggered differently, as
explained later). Calls come in from the PSTN via a 3745 gateway that
is also running the gatekeeper, CCM is registered to it via a gatekeeper
trunk and the internal destinations are routed to RAS. All other call
flows are seemingly undisturbed, especially normal call clearing does
work as expected. I tracked this issue down to:

1) SETUP comes in from PSTN and proceeds with allocated B-channel:

Dec  7 10:18:53.646 CET: ISDN Se2/0:15 Q931: RX <- SETUP pd = 8  callref = 0x0E8A 
        Sending Complete 
        Bearer Capability i = 0x8090A3 
                Standard = CCITT 
                Transer Capability = Speech  
                Transfer Mode = Circuit 
                Transfer Rate = 64 kbit/s 
        Channel ID i = 0xA9839A 
                Exclusive, Channel 26 
        Calling Party Number i = 0x2181, '3514716466' 
                Plan:ISDN, Type:National 
        Called Party Number i = 0xC1, '2588999' 
                Plan:ISDN, Type:Subscriber(local) 
        High Layer Compat i = 0x9181

Dec  7 10:18:53.658 CET: ISDN Se2/0:15 Q931: TX -> CALL_PROC pd = 8  callref = 0x8E8A 
        Channel ID i = 0xA9839A 
                Exclusive, Channel 26

2) Gateway relays the SETUP to the CCM via H225:

  Q931 Message IE Decodes
Protocol Discriminator : 0x08
CRV Length             : 2
CRV Value              : 0x3ADB
Message Type           : 0x05: SETUP
 Bearer Capability: Length Of IE=3
 Data 8090A3
 Calling Party Number: Length Of IE=10
 Data 41813034373136343636
 Called Party Number: Length Of IE=4
 Data C1393939
 User-User: Length Of IE=275
 [...]

3) Gateway receives call proceeding from CCM:

  Q931 Message IE Decodes
Protocol Discriminator : 0x08
CRV Length             : 2
CRV Value              : 0xBADB
Message Type           : 0x02: CALL_PROC
 User-User: Length Of IE=51

4) Gateway receives release complete with cause code "User busy" from CCM:

  Q931 Message IE Decodes
Protocol Discriminator : 0x08
CRV Length             : 2
CRV Value              : 0xBADB
Message Type           : 0x5A: RELEASE_COMP
 Cause: Length Of IE=2
 Data 8091
 User-User: Length Of IE=33

All this happens almost instantaneously and is what is to expect in this
case. The problem now is the gateway does not act upon the release
complete cause user busy for the next 15s (which is probably some timer
running out), it simply leaves the call dangling.

After these 15s or so, the gateway sends a disconnect to the PSTN,
bearing an interesting cause code:

5) Gateway disconnects the call on the PSTN side:

Dec  7 10:19:08.683 CET: ISDN Se2/0:15 Q931: TX -> DISCONNECT pd = 8  callref = 0x8E8A 
        Cause i = 0x80A6 - Network out of order
Dec  7 10:19:08.735 CET: ISDN Se2/0:15 Q931: RX <- RELEASE pd = 8  callref = 0x0E8A
Dec  7 10:19:08.739 CET: ISDN Se2/0:15 Q931: TX -> RELEASE_COMP pd = 8  callref = 0x8E8A

Oops. Network out of order? How this?

IMO that's a bug on the gateway, maybe triggered by some change in
signaling of user busy by CCM 4.1.3 SR4b. I'm just completely baffled
on what to do next. Anyone seen something like this before or has a
hint on the matter?

TIA,
Andre. 
-- 
                  The _S_anta _C_laus _O_peration
  or "how to turn a complete illusion into a neverending money source"

-> Andre Beck    +++ ABP-RIPE +++    IBH Prof. Dr. Horn GmbH, Dresden <-