CallManager 4.1(3) One-way voice problem

[jimieee802]

Environment: One MCS CallManager at 4.1(3) in test mode using all VoIP 7960 phones powered by 3560 POE switch. There is no Voice gateway or PBX involved to date. just
on-net calling. I have not had the time to keep up with the latest maintenance including Phone LOAD firmware & must now await a new SmartNet contract to obtain any needed
maintenance. Additionally each IP Phone has the following: App Load ID: P00308000300 Boot Load ID: PC03A300 ver 8.0(3.0). The Call Manager is at the following versions
Cisco CallManager System version: 4.1(3)sr3b
Cisco CallManager Administrator version: 4.1(0.11)
Cisco CallManager Installation ID: 4.1(3)sr3b

Problem: Someone calls me or I call them (both 7960 phones) after off-hook the call status is "connected" on the phone display. However, only one voice can be heard. If I or the other phone dials back a normal conversation takes place. This can happen either way with any phone. All phones are Vlan tagged and are leasing the proper IP address on the Voice Vlan subnet. The TFTP service is on the single CallManager / Publisher. All local phones are on the same 3560 switch (POE) at version 12.2(25)SEB1. CallManager
is local to the building.and on the same subnet. There should be no router involved in the connection. A CISCO document that I am reading (chapter title = 'Solving Voice
Problems' states: "One-way audio - a conversation between two people in which only one person can hear anything - is not actually a voice quality issue, but it may be experienced on a VoIP network." But the document generally assumes a Gateway is involved which is not our environemnt. Since the two phones are CONNECTED I am assuming that the CallManager is not in the picture since call setup has completed. Silence Suppression is set to False. Any ideas?

 

[ADB100]

One way speach is usually an IP routing issue - i.e. there is a path one way between the one IP Phone and the other, but not a return path. Verify the IP routing path both ways from an end device in each of the VLANs where the phones are.

Have you tried a call between two phones on the same switch and in the same Voice VLAN?

HTH

Andy

 

[MG2005]

This has good info on call setup/behavior between 2 IP phones reg'd to the same CM cluster :

http://www.cisco.com/en/US/products...leshooting_guide_chapter09186a00803fa951.html

You need to sniff and/or trace a bad call and look for this I believe:

Cisco CallManager uses the following messages to order the station to begin transmitting the audio and video streams to the indicated remote Cisco IP Phone IP address and port number.

16:05:45.265 CCM|StationD - stationOutputStartMediaTransmission
tcpHandle=0x4fbbc30 myIP: e74610ac (172.16.70.231)

16:05:45.265 CCM|StationD - RemoteIpAddr: e64610ac (172.16.70.230)
RemoteRtpPortNumber: 17054 msecPacketSize: 20
compressionType4)Media_Payload_G711Ulaw64k


16:03:25.328 CCM|StationD(1): TCPPid=[1.100.117.1]
OpenMultiReceiveChannel conferenceID=16777217 passThruPartyID=1000011
compressionType=101(Media_Payload_H263) qualifierIn=?. myIP:
e98e6b80 (128.107.142.233)|<CT::1,100,11,1.1><IP::><DEV::>


16:03:25.375 CCM|StationInit: TCPPid=[1.100.117.1]
StationOpenMultiMediaReceiveChannelAck Status=0, IpAddr=0xe98e6b80,
Port=65346,
PartyID=16777233|<CT::1,100,105,1.215><IP::128.107.142.233>


16:03:25.375 CCM|StationD(2): TCPPid = [1.100.117.2]
star_StationOutputStartMultiMediaTransmission conferenceID=16777218
passThruPartyID=16777250 remoteIpAddress=e98e6b80(66.255.0.0)
remotePortNumber=65346 compressType=101(Media_Payload_H263)
qualifierOut=?. myIP: e98e6b80
(128.107.142.233)|<CT::1,100,105,1.215><IP::128.107.142.233>


In the following traces, the previously explained messages are sent to the called party. The messages that indicate that the RTP media stream has been started between the called and calling party, follow these messages.

16:05:45.312 CCM|StationD - stationOutputStartMediaTransmission
tcpHandle=0x4fbb150 myIP: e64610ac (172.16.70.230)

16:05:45.328 CCM|StationD - RemoteIpAddr: e74610ac (172.16.70.231)
RemoteRtpPortNumber: 18448 msecPacketSize: 20
compressionType4)Media_Payload_G711Ulaw64k

16:05:46.203 CCM|StationInit - InboundStim - OnHookMessageID
tcpHandle=0x4fbbc30


The calling party Cisco IP Phone finally goes on hook, which terminates all the control messages between the Skinny Station and Cisco CallManager as well as the RTP stream between Skinny Stations.

16:05:46.203 CCM|StationInit - InboundStim - OnHookMessageID
tcpHandle=0x4fbbc30

 

[jimieee802]

ADB100 TechUser:
This morning I configured a port on the 3560 switch for the same VOIP Vlan # & was able to ping the IP addresses of both VOIP phones & also the CallManager (no routing hops etc). This problem also occurs when calling from a remote building (still OnNet) where the calling phone can not hear but when calling back immediately both parties can have a successful conversation.

MG2005 TechUser:
I will need to run those sniffer traces when I return to work on Tuesday February 20. I will review the Cisco document this weekend.

I was wondering what the recommended switch port configuration would be for the CallManager itself. Also are there special configuraton recommendations for the up / down link ports on the distribution to access (3560) switches?

Thanks for both of your suggestions. Have a great weekend.