[Asterisk-Dev] DMS-500 CID NAME Problem

[Matt Fredrickson]

On Wed, Jun 22, 2005 at 09:17:59PM -0500, JR Richardson wrote:
> I have several * servers connected to T1 PRI's from various service
> providers in multiple locations the US.  All the * servers use the same
> hardware with the same OS and * version CVS-v1-0-11/09/04-12:27:27.  When
> connected to 5ESS Switches, using the NI2 (national) PRI protocol, the CID
> name and number come across fine and populate into the * CDR fine.  I
> connected to a DMS-500, NI2 (national) protocol and the CID name doesn't get
> populated in the * CDR.  The only variance in the PRI debug outputs is this,
> from Asterisk debug when connected to the DMS switch:
> 
> Jun  9 16:41:15 WARNING[30369]: chan_zap.c:7133 zt_pri_error: PRI: !!
> Facility message shorter than 14 bytes

That's an old warning message.  You should probably be using CVS-HEAD if you're
having any issues with facility related messages.

> Facility (len=12, codeset=0) [ 0x9f, 0x8b, 0x01, 0x00, 0xa1, 0x01, 'H',
> 0x02, 0x01, 0x02 ]

I don't have any clue what is in that message.  There is no caller name
information encoded in that facility IE.

> Can someone explain what len=12 means?  On all the good results from the
> other switches, the len=23 or greater.

It just means that facility IE is only 12 octets long.

> The interesting part is that the CID name does come into the * server and is
> forwarded to the destination phone, the CID name and number does come across
> from the service provider.
> 
> I'm thinking there is something particular with this DMS-500 causing this.
> Is there a special terminating setting I can throw into Zapata.conf to
> compensate for this short facility error?
> 
> I was able to test with another DMS-500 with no problems, so I don't think
> it's a Nortel issue, I believe it to be a configuration problem on the
> serving switch PRI or possibly farther upstream in how they are gleaning the
> CID info from the SS7 stream.
> 
> I can send full debug outputs from both working and non-working servers if
> needed.

If you can put a full trace out from the switch that doesn't work, then we can
see if it contains the caller name information.  From what you've posted, I haven't
seen any thus far.

Matthew Fredrickson