OK. I'll have to open a case when I have some time to commit. I've written down a few numbers from TAC engineers and they're willing to help reproduce the problem. I think most of them were from 919 to tell you the truth. But I can't remember. I'll post whatever I find out from the TAC to the list.
----- Original Message -----
From: Wes Sisk
To: Lelio Fulgenzi
Cc: cisco-voip@puck.nether.net
Sent: Tuesday, August 10, 2004 10:42 PM
Subject: Re: [cisco-voip] First character missing from calling name ID on some calls from Cisco TAC engineers
Hi Lelio,
Yep, you've got MGCP. Hmm, CSCdu54048 should have taken care of this for you. We're gonna need that TAC case. Please include a CCM trace that shows an inbound call setup from a Cisco employee. Have you noticed coming from any specific Cisco campus/area code? RTP(392) SJ(408)...?
CSCdu54048
When CSCdu18827 was resolved in CallManager 3.1(0.163), it was discovered
that a number of ISDN switches implementing the DMS PRI protocols do not
adhere to Nortel's specification for the Display IE.
Specifically, there is an "extra" leading character that Nortel sends at
the front of the Display IE that should not be used as part of the
actual display for the telephone set.
The problem is that now, all PRI gateways in CallManager that are configured
to use one of the DMS PRI protocols assume that this extra character is there.
If the ISDN switch is not sending this extra leading character, and rather
the first character is part of the display name, then the first real
character of the display name will be discarded from the telephone display
on the IP phone. This can happen in any of the messages where the Display IE
can be carried (e.g. SETUP, CONNECT, etc)
For example, if I call from the IP phone on 3.1(0.163) to another phone
through the gateway where my name display should be John Doe then when
the call connects and the CallManager gets Display IE = John Doe then
what actually gets displayed is:
ohn Doe
This is being filed as an enhancement so that for any time we receive a
Display IE on the PRI gateway configured with one of the DMS PRI protocols,
we should look at the most significant bit in the byte following the length
field in the Display IE.
If it is an ASCII character, then we do nothing and continue normally. If
it is a non-ASCII character, then we strip off the entire byte before handing
off the name display to the destination device.
Lelio Fulgenzi wrote:
Hi Wes. We are using 6608 blades in a 6500 chasis, so it's not IOS - it's CatOS, I think the bugs were for IOS in routers. I'm pretty sure they are running MGCP, since I was told if it has the T1 icon, it's MGCP, and it's H323 if it has H323 in the icon (all in the gateway configuration screen on CallManager v3.3(3)sr4a).
It's certainly wierd - it only occurs with calls from Cisco TAC staff.
----- -----
Lelio Fulgenzi, B.A. lelio@uoguelph.ca.eh
Network Analyst (CCS)
University of Guelph FAX:(519) 767-1060 JNHN
Guelph, Ontario N1G 2W1 TEL:(519) 824-4120 x56354
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
remove the 1st letter of the canadian alphabet from my email, eh!
----- Original Message -----
From: Wes Sisk
To: Lelio Fulgenzi
Cc: cisco-voip@puck.nether.net
Sent: Tuesday, August 10, 2004 8:08 PM
Subject: Re: [cisco-voip] First character missing from calling name ID on some calls from Cisco TAC engineers
Hi Lelio,
Are you using h323 gateways?
CSCea76050, CSCdz86750,
In the Avaya implementation of DMS trunks, they do not support
the special leading character. The Avaya PBX's are configured
for DMS switch type, and for that configuration the Avaya PBX
does not insert(on outbound) or strip(on inbound) the special
leading character.
The possibilities:
1) some endpoints do support the DMS special leading character
on trunks configured for the DMS protocol (Nortel PBX's)
2) some endpoints do not support the DMS special leading character
on trunks confiugred for the DMS protocol (Avaya,Cisco,other)
3) it is not possibile for intermediate VOIP hops to modify DisplayIE
4) it is possbile for intermediate POTS hops to modify the DisplayIE
Basic opertion:
Default IOS behavior has to be maintained. The display IE should be passed
from pots->voip and from voip->pots without modification. The DisplayIE should
not be removed and inserted in the UserUserIE RAW data.
What is needed to make this work for customer EY:
1. DO NOT MODIFY DisplayIE (transparently pass DisplayIE without modification in all cases)
What is needed to make this work for all possible customer permutations:
1. Do nothing (blindly pass DisplayIE, make this default just like in previous IOS versions)
2. CLI to stip leading char going pots->voip
3. CLI to insert leading char going pots->voip
4. CLI to insert leading char going voip->pots
5. CLI to strip leading char going voip->pots
Leading character "0xB1" for DMS 100, or "0xB2" for DSM_250.
********************
We have confirmed that the originating Jax CCM did not insert the 0xB1, but the 0xB1 is present when
the call arrives at NY.
I believe this is due to the IOS request CSCdx12421. This was initially requested to have a hidden
CLI to add/remove the 0x81 character. However, the final implementation in is a hard coded insertion of
0xB1 when it is not initially present.
Please confirm the version of code on the Jax 36xx router.
We need to confirm once more that the 0xB1 is not in the Dislpay when it leaves the Jax CCM, but is
there in the Debug isdn q931 when the call leaves the Jax router.
Going forward we have 2 options:
1. use MGCP gateways everywhere. Callmanager still allows the option of inserting 0xB1 or not
inserting it,
We could configure MGCP IOS gateways in Troy, Jax, and Hou?? to avoid this issue.
2. upgrade CCM to at least 3.2(2c)spF everywhere and enabled the parameter
"SendExtraLeadingCharacterInDisplayIE"
everywhere. CCM requires this flag be set in the gateway configuration page to check for the 0xB1
character and remove it.
With this box checked, incoming messages from MGCP gateways will be checked and stripped of a
leading 0xB1 character.
However, this will also have CM send the leading character out as well, so the far end CM will have
to expect this and remove it.
Thus the need for 3.2(2c)spF and this configuration in all locations.
/Wes
Lelio Fulgenzi wrote:
> In order to get calling name ID working properly, we had to check off
> the following parameter in our *Digital Access PRI* gateway:
>
> [x] Send Extra Leading Character In DisplayIE***
>
> Once we set that parameter, the special character that appeared before
> the calling name (usually an ASCII graphic character) went away, and
> all was well.
>
> However, now, when I get a call from some Cisco TAC engineers (this is
> the only time I've seen it), the first letter of their name is
> missing. Instead of "Tom Smith", I see "om Smith".
>
> Has anyone else run into this?
>
> Just thought I'd send a note out before I open a case with the TAC on
> this. I'm worried that between some CallManager-to-CallManager
> communications, this parameter is actually causing the problem and if
> it's something that can be avoided if we decide to go with some
> off-site CCM installations.
>
> Lelio
>
> --------------------------------------------------------------------------------
> Lelio Fulgenzi, B.A.
> Network Analyst (CCS) * University of Guelph * Guelph, Ontario N1G 2W1
> (519) 824-4120 x56354 (519) 767-1060 FAX (JNHN)
>
> "This signature may contain traces of nuts"
> --------------------------------------------------------------------------------
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>cisco-voip mailing list
>cisco-voip@puck.nether.net
>https://puck.nether.net/mailman/listinfo/cisco-voip
>
>