Multiple Device Pools, Single SIP Trunk

[IT Service Desk]

Hi, I was hoping someone can shed some light on this. I am not an IPT engineer.  We are putting in a new CUCM 10 but the person who started this is no longer here and am just trying to pick up the pieces left behind.  Thank you in advance.

Question -

Ignoring for now any E.911 stuff... Can 2 device pools make outbound calling against a single SIP trunk (1 IP address)?  Office1 has a pub/sub and a single SIP trunk. Office2 has neither. We wanted two separate device pools. But the SIP trunk (found under Device | Trunk) can only be associated with a single DP.

We use an MPLS network dedicated to voice. The service provider is able to force inbound calls for both area codes to that SIP trunk.  I guess my problem is on the outbound.

If it is possible, how do we make this happen?  If not, does that imply that both offices must be in the same DP?

Thanks.

[Brandon Pierce]

Device Pools only affect where traffic gets routed if you are using Standard Local Route Groups.  If that is the case, then whatever is first in that route group will take precedence unless you are using circular routing.  So yes, you could have 50 device pools and they could all route out a SIP trunk with no issue as long as the route pattern they catch sends them in that direction.  The other way is to not use Standard Local Route Groups and route based on matched pattern straight to whatever destination is set.

[IT Service Desk]

Brandon:

Thank you. If I was an IPT engineer, it might make better sense but I take your explanation as correct. :-) What did you mean by "If that is the case (that is, if I'm using Std Local RG), whatever is first in that route group will take precedence..."  I think I am using Std Local RG (see below) but I don't see how the presence of this causes all outbound voice traffic to go out the SIP trunk.  The only top/down list I see is within a route list, but the Std Local Route Group is listed second.

Under Route Group

I have a single RG1 listed and the properties of it show my SIP trunk added as the only available device and and only RG member.

Local Route Group Names

I see we do have a Standard Local RG.

Route List

I see something he created called RL-SLRG. Within that definition, My RG1 is on top and RL-SLRG is listed second.

Route Pattern

When I view the ones of interest, I see that the Gateway/Route List is specified as the RL-SLRG.  For example, if you dial 9,1,area+number, use the RL-SLRG it says.  How does this even work to know that RL-SLRG is to use my service provider's SIP trunk for outbound?  I don't see thru anything on the web admin interface that defines this.

I don't think the forum was intended to get a complete set of steps on how to implement so I won't bug you like that. If you can kindly give me some guidance on the above, the significance of a SLRG, etc., we can try to take it from there. Thank you again.

[Brandon Pierce]

In the device pool of the phones, there is an option for Standard Local Route Group.  Is that set to something specific?  If it is, you need to check that Route Group to see where it is sending calls as that will be your egress point upon a Route Pattern match.  Hope that makes sense?  Top/down indicates it will always use the top one until it is either loaded or not available which is usually desired unless you for some reason want to round robin the calls with a circular algorithm.

[IT Service Desk]

Hi Brandon, sorry I hit "Report" by accident instead of Reply. I did an Unreport. :-)

Within both Device Pools I see "Local Route Group Settings" and the Standard Local Route Group is set to < None > on both.

Sorry, hopefully last set of questions and will mark this as answered/correct.

1. Given the Device Pools show < None > for the above, does this mean, while the guy defined some SLRG, that I'm not actually using them for Device Pools?

2. Given that my Route List shows my RG1 as the top, does that mean it gets processed first (before the SLRG)?  If so, does that effectively mean I'm not really using SLRG?

3. But my Route Pattern shows that outbound calls should go thru the SLRG.

Sorry, very confused. I wish this were an ASA or something. Just easier for me... 

Thank you.

[Brandon Pierce]

If the device pools are set to "None" then they will route with whatever the Route Pattern has set.  So if the Route Pattern has XXXXX_RL then you need to check the route lists.  If the Route Pattern is set to XXXX_Trunk or Gateway then it goes straight out that gateway or Trunk.  Can you post some screenshots of the setup?  This looks like a basic design from what you have said.

[IT Service Desk]

p.s. thanks for this...

I attached a screenshots of one DP, a route pattern for an outbound call, the route group and route list.

[Brandon Pierce]

Ok with that given, NY_SIP is the device it is using to egress your network.  Not sure why there was a label of SLRG in there when it isn't an SLRG. So any device that makes a call without an Standard Local Route Group in their device pool will use the route pattern's target.

[IT Service Desk]

It sounds like you have an answer to which I didn't understand... With this particular setup: 

1. A phone in DP-Office2 makes an outbound call. It will go via route pattern match, which currently says "use this particular route list".
2. The route list says (top/down), first one to try is RG1 route group.
3. RG1 route group is using the NY_SIP. 
4. Unless I missed your explanations badly, it seems like then the existing setup should work but the phones in this DP get a fast busy when making the outbound external call.  The only thing I can see is that the NY_SIP properties are homed to DP-Office1. 

[Brandon Pierce]

1. Are these two offices in different locations or just marked differently to distinguish a section? 
2. Do the DP2 phones register to CUCM? 
3. If the DP-Office2 phones are up and running and at a remote site, have the calls ever been working?
4. Does the second DP have "None" as a SLRG? Your previous statement led me to believe it is set to "None"
5. Are these phones SCCP or SIP registered?  You could look at the logs from RTMT and see what the issue is.  RTMT is a side program under plugins that can let you pull detailed logs as to why things are happening.
6. Are there any calling restrictions like CSS's and Partitions getting in the way?  The previous engineer may have set the block message to a fast busy seeing how the system is already designed.

Getting off here soon so someone else may need to keep this discussion going. Please don't forget to rate all useful posts and mark a correct answer should one be provided.

[IT Service Desk]

1. Two different locations both connected via private point-to-point.

2. DP2 phones (office2) can register into CUCM in office1.

3. Outbound to outside numbers have never worked out of DP2/office2.  Interoffice works. Inbound works as well. The provider is able to have both area codes come in thru that single circuit in the main location.

4. Yes, both DP have "none" for SLRG.

5. Mix, strangely. Most are SIP. A few stragglers on SCCP.

6. No known restrictions under CSS/Partitions that I can see. 

Will try to run the RTMT.  Brandon, you are way cool. Let me try this and get back to you. If not tonight, tomorrow.

[Brandon Pierce]

Are you running pure G.711 across the WAN link or are you forcing G.729?  This would dump your call if you arn't transcoding properly from one codec to another.  you may be pure G.729 between the links but not transcoding hitting the PSTN to G.711 is my thought.  Almost sounds like a media resource problem at this point.

[IT Service Desk]

I ran RTMT and assuming I did it correctly, I saw the failed test call and it says "( 47 ) Resource, unavailable, unspecified. Explanation: A resource unavailable event occurred and no other cause in the resource unavailable class applies".

RE: your other post, when I look at Regions, the only mention is the Max Audio Bit Rate which shows 64 kbps (G.722, G.711). This is consistent across both regions.  The preferred codec is to use the system default of Factory default low loss.

Not sure about media resources.  I don't understand the naming convention of appending _2 or _4 after each resource but I see a pair for:

• ANN
• CFB
• MTP
• MOH.

For each resource,

• "Resource_2" is homed to DP2
• "Resource_4" is homed to DP1.  
• "Resource_2" shows the IP of the publisher.
• "Resource_4" shows the IP of the sub. But both are registered against the sub, if that makes sense.

There is an MRG for each DP.  For example:

• MRG_ANN_DP2 (with resource ANN_4)
• MRG_ANN_DP2 (with ANN_2).
• Etc.

Then there are 2 MRGL's - one for each DP.  Each MRGL though contains all the MRGs.  

Is my problem a media resource issue?

[Brandon Pierce]

I knew it, media resources are hosing you up.  When you make a call from office2, what is the path?  If you are G.711 all the way across you don't need transcoding but something else is hosing you up. 

1. ANN is annunciator which is responsible for call progress tones, the "Cannot complete this call as dialed", etc.
2. CFB is Conferences on the CUCM, can be either hardware or software, software is what comes with CUCM and is capable of g.711u/a and wideband codecs
3. MTPs are your hold/resume abilities in which CUCM can hold the call, transfer it, etc.  Hardware MTPs can also be used beyond what the CUCM can handle.  MTPs basically are the bridge that allow media streams to be setup and taken down separately.
4. MOH is just music on hold

Cause code 47 is definitely a media resource issue.  So you checked every region and nothing noted G.729 anywhere? There is something called a Dialed number analyzer  in CUCM as will show the call path from the perspective of a phone.  When you log into CUCM

1. Upper right corner is a dropdown go to Servicability
2. Tools --> Dialed Number Analyzer
3. Analysis --> Phones
4. find a phone at site 2 and select a line on that phone once you click the phone
5. Type in a valid PSTN number that the phone can dial(permissions allowing of course)
6. See what the call flow looks like just to verify the path it is taking is correct

Other than that, I would need to know the topology and how everything is laid out with the DNA tool and Regions.  My next question is, under the System--> Locations, you don't have anything setup for the sites there do you?  If you do, are those put on any of the phones?  Default is Hub_None and Phantom which I doubt he changed if that is all that is on there.

I bet if you completely removed all MRGs, MRGLs, and their assignments the call would complete assuming G.711 is truely what is being used.  In the RTMT logs for the call, on a SIP phone, what is the SDP sending the initial INVITE with assuming you are using early offer?  SDP is Session Description Protocol which looks like this:

Initial SIP INVITE message for call setup:

INVITE sip:13@10.10.1.13 SIP/2.0
Via: SIP/2.0/UDP 10.10.1.99:5060;branch=z9hG4bK343bf628;rport
From: "Test 15" <sip:15@10.10.1.99>;tag=as58f4201b
To: <sip:13@10.10.1.13>
Contact: <sip:15@10.10.1.99>
Call-ID: 326371826c80e17e6cf6c29861eb2933@10.10.1.99
CSeq: 102 INVITE
User-Agent: Asterisk PBX
Max-Forwards: 70
Date: Wed, 06 Dec 2009 14:12:45 GMT
Allow: INVITE, ACK, CANCEL, OPTIONS, BYE, REFER,
  SUBSCRIBE, NOTIFY
Supported: replaces
Content-Type: application/sdp
Content-Length: 258

Then the SDP:

v=0
o=root 1821 1821 IN IP4 10.10.1.99
s=session
c=IN IP4 10.10.1.99
t=0 0
m=audio 11424 RTP/AVP 0 8 101
a=rtpmap:0 PCMU/8000
a=rtpmap:8 PCMA/8000
a=rtpmap:101 telephone-event/8000
a=fmtp:101 0-16
a=silenceSupp:off - - - -
a=ptime:20
a=sendrecv

Note the PCMU/8000, in the US we generally use PCMU for everything.  If you arn't offering this and there isn't anything in there when site 2 calls, that is part of the issue, but I suspect this isn't a SIP trunking issue and again, purely a media resource problem.  Once I get a better picture of your topology, we can go from there.  I have a feeling if you set the MRGL on a phone from Site 2 to an MRGL of Site 1, you call would complete.

[IT Service Desk]

Thank you for the primer on the various media resources. It was very good information for me.  Sorry lengthy post but you had a bunch of good questions.

Call Path

When I make a call from office2 (NJ), the call traverses our 100 Mbps point-to-point into office1 (NY). Point-to-point util is no greater than 10% (peak over 6 weeks) and average util of about 3%. NY has the pub/sub and voice circuit. The NJ call is supposed to go out that way.  Pretty vanilla.  When office2 is on its own DP, it fails. But if set to the office1 DP, it works well.

DNA Tool

I ran the tool twice. First using a working phone from office1. I then compared that to the results of using a non-working phone from office2.  The results are an identical match.

Under Call Flow, the first section is the Translation Pattern section and it doesn't show much but appears OK. Just some stuff about transformations and such.

Route Pattern looks like it just says it matches the 9.1[2-0]XX[2-9]XXXXXX. 

Rout List shows 2 Route Groups. First one just says RG1 and to use the End Device Name NY_SIP.  The second one says SLRG is the route group name but nothing more really listed there other than some transformation information.

Next, confirming I only have 2 regions (corresponding to offices 1 and 2) and neither have G.729 in them.

Topology:

• Basically office1 is NY. It has a pub/sub and the provider's SIP.  This is MPLS network dedicated to our voice only. Provider can take both 212 (NY) and 201 (NJ) area codes and permit inbound on that SIP and do outbound as well for both area codes.
• Office2 is NJ but currently has no voice infrastructure. Just need to get these 7965 phones working thru the NY infrastructure for now until we get the other equipment and circuits ordered for NJ.
• Offices 1 and 2 connected for data via a 100 Mbps point-to-point circuit.

System > Locations:

• Two locations setup - one for NY and one for NJ. Both setup identically.
• NY has NJ as its adjacent location.  In there it shows NJ having a weight of 50, audio b/w 48, video b/w 384, immersive b/s 384.
• Under Advanced, audio, video and immersive have Unlimited chosen.
• Location RSVP settings show both NJ and Phantom with no reservations.
• NJ setup identically but in reverse.  That is, it just shows NY as the adjacent partner.
• There is no Hub_None. The other 2 locations are Phantom and Shadow. Neither of these have any adjacent locations listed.

System > Regions:

• 2 regions - 1 for each office.
• Looks as if they are all using a system default. Codec preference factory default low loss.
• Max audio bit rate 64 kbps (G.722, G.711)
• Max session bit rate for video 384 kbps
• Max session bit rate for immersive 2147483647 kbps
• Under modify relationship to other regions - all settings are set to keep current settings.

The RTMT thing from yesterday that had that 47 Resource unavailable lengthy error.  When I double-click it, it shows a call flow diagram of:

1. Invite
2. 100 trying
3. 183 Session Progress
4. CANCEL
5. 200 OK
6. 487 Request Cancelled
7. ACK

The next tab of Message Details\Detailed SIP\Log File just shows that request was canceled and reason same as yesterday of Q.850; cause=47. 

Thank you.