MOH failed to Unicast, calls on HOLD dropped

[markhayek]

We have 2 MOH virtual servers configured to source from 2 different TFTP servers in the cluster. The MRGL for MOH on the SIP trunks for outbound, off-net calls use that MRGL which lists both MOH servers in sequence.   Whenever a call used the MOH #1 , the call heard music, and was OK on HOLD. But when a call hit MOH #2, there was NO MUSIC playing, and the call on HOLD dropped within 8 seconds since the SIP SDP showed inactive RTP stream, so Verizon dropped the call.  Once we directed all MOH to MOH #1, then reset the MOH server #2 and reset the MRGL list, all MOh worked again normally.  There was no evidence of a problem, but  looking back, I see this alert in the TFTP server syslog,

Jul 16 15:32:54.615 UTC : %CCM_MEDIA_STREAMING_APP-IPVMS-3-kMOHTFTPGoRequestFailed: Start of download request failure. File Name: Cluster ID:C4 Node ID:CLT-UCMTFTP1C4

Could this be the cause of the MOH to stop playing , and result in silence on the line ( no RTP) ?  Is this a known problem in CUCM 6.1.5 version ?

[Anonymous]

I am going to assume MOH #1 was already working at some point and then it stopped working.

If that is the case, I could not find any known issues which describe the problem you faced.  I would suggest turning on the IPVoice Media Streaming App, and CCM traces to detailed level on all nodes.  If the problem reoccurrs provide a set of IPVMS and CCM traces from a working to a non working period to Cisco TAC.  If the problem doesn't reoccur after sometime, you can set the trace level back to a default level.

This document describes setting and collecting CCM traces.

http://www.cisco.com/en/US/products/sw/voicesw/ps556/products_tech_note09186a0080094e89.shtml

Once again, you'll want to do this entire procedure for CCM (Cisco CallManager) and IPVMS (IP Voice Media Streaming App) traces.

HTH

[markhayek]

Good idea, IPVMS traces set to detailed, but only have 16 files of 1 Mb each, increased files to capture this event if it recurs.