I have a UC520 with two teleworkers in different sites. All sites connect back to UC520. Calls between phones and the "Hub" work well but calling between each other is not working. These are true L2L ipsec tunnels with no mesh. Do I need to have ip connetivity directly between phones in order for them to work? I had always thought as long as phone could hit CME the magical world of call delivery would work.
This might be a silly questions but do you have transcoding in place?
Are you forcing any particular codec on the teleworker phones on the CME side?
Not a silly question :) there is no transcoding configured on the system, all G711 to make things simple. I have this problem now with two clients. One using traditional CME on ISR and the other with UC500. Both deployments are setup with a hub and spoke topology. Static site-to-site (non EZVPN) tunnels back to the head end. Teleworkers can call great to main office phones. But the problem is teleworker to teleworker. Teleworker tunnels are setup to just encrypt data to head end and not between sites (like a DMVPN ..etc / Mesh) so I am thinking this has something to do with it.
I wonder if its ACLs on eack teleworker router to allow the subnet of the other?
When you set up a teleworked with CCA, it takes care of TW Router and UC500, but it doesnt know about the others, maybe? Not sure. Thinking out loud I guess...
I can confirm that they do not have acl's for the other "spoke" subnet. So the big question is: On a CME system must all individual SCCP phones have IP access each other in order to make successful calls or can you hairpin RTP ..etc for voice through the router.
What happens when the two remote Teleworkers attempt to call each other. By default RTP traffic is directly between the phones and control traffic is through the central unit. You can configure hairpin in CME but I would assume that would be your issue if the two phones can call each other (ringing) but are unable to connect the audio.
That makes sense to me what is happening is in this scenario is the user can call the other phone and make it ring just fine. However this is no sound at all through the channel.
However on our other system we have a lot of teleworkers, configured the same and there is really no problem between teleworkers except for one single user. Everyone else is just fine and we have the same tunnel configuration where we are only encrypting for the head end office subnet.
What would be the command to allow for hairpinning media through cme? Something under ephone ?
Ok this issue has been resolved. The problem was that I did not have the "mtp" command under the ephone's in question. Without this command in place the phones will attempt to RTP directly between phones. If this command is enabled it forces RTP through the CME router. Thank you all for the brainstorming to figure this out.
Thanks Kenny, I will look into this.
Technical Solutions Architect
Small Business Sales
Cisco Systems Inc.
Research Triangle Park
North Carolina, U.S.A
Most recently I used the new 2.2(4) - We had UC520 where we configured the "mtp" com
mand manually under each ephone. We then (using 2.2(4)) Upgraded the box to 8.0.2 . Once the upgraded was complete did some configuration of new phones using CCA 2.2(4) . The next day users told me the problem had came back. I checked the ephones (these were existing ephones not the new ones we added) and they did not have the mtp command. I put the command back in and resolved the issue.
I just heard back from the CCA Development team. I am sorry you found this issue.
Seems like the dilemma is we dont know which phones are remote and which are local in CCA.
Having this command for all phones is inefficient, because this would mean that calls (RTP) between local IP phones will have to go through the UC500. This is usually reserved just for remote teleworker phones.
We have a CDETS open for this:
CSCsm36527 Assigned / 851 voice remote teleworker, 'mtp' need to be configured under ephone-dn
So while the issue was found and we want to fix in 2.2.5, the question is how...
...we maybe can expose the MTP option in the User Inteface of CCA for you to select....
The logic we have in CCA is always send down 'no mtp' when phone type is not "CIPC" (soft phone).
Please hang in there with us while we figure it out...
Steve I like the idea of having the mtp as a selectable option under the configuration of a phone. Alternativly I wonder if you could detect when a phone registers to CME from a subnet which is different from the ip subnet configured on the vlan interface of the voice vlan. That would would be a great indicator of remote or local. I am not sure how this would fair with an EasyVPN teleworker scenario but its a thought. In the end as CCA does not eat my configuration if I configured it manually through the CLI I will be happy
apologies for waking an old thread but in searching for an answer to my critical client issue i stumbled onto this great exchange (nice job guys on this one BTW love the collaboration)
So i have a uc560 in NYC and a nortel cs1000 in moscow
cisco 3xxx in moscow <> uc560 in NY
SIP trunk between the two (CLI using dialpeers)
VPN tunnel between the two as well
calls between sites work great
cisco to nortel (tdm phone)= good
cisco to nortel "pc client"= good
nortel pc client to cisco = good
nortel tdm to cisco = good
data/routing to each LAN= good
both nortel phone types to unity VM = good
toll bypass calls pc client to cisco UC to pstn (isdn PRI) = good
toll bypass calls nortel tdm to cisco UC to pstn (isdn PRI) = NO GOOD
call flow = moscow nortel dials xx2125551212 > xx steering code sends
call over SIP trunk > nortel strips xx and sends 10 digits to NY uc560 > call hits UC560 and dial peer for pstn (no "9" or access code needed) > call routes out ISDN PRI & RINGS local NY phone > I answer call > no audio in either direction but call is connected
so must be an MTP or Codec/transcoding issue ?
now I have hardware conference setup but no other DSP resources allocated by me (just whatever is created by CCA for PRi and HW conf)
odd that nortel pc client works but tdm endpoint does not - both use same nortel signaling/sip server but pc client must use different DSP resources i guess?
anyway i have debugs and configs if helpful. time sensitive issue, please advise
Sent from Cisco Technical Support iPhone App
cant seem to attach a config?
so issue is call connects but no audio in either direction
SIP trunk > uc560 > isdn PRi
moscow side = nortel cs1000 for voice and SIP trunk +
cisco 3xxx (3700 ?) for site to site vpn tunnel/edge router
NYC = UC560 for site tunnel and SIP trunk + PRI for pstn voice
Fios internet 150mb w/static IP's
Sent from Cisco Technical Support iPhone App