Duo Video sending drops out after 30 seconds

[damienjones1]

Hi, I am experiencing an intermittent issue with Duo video content sending dropping out unexpectedly about 30 seconds after it is being started. 

The systems systems involved are comprised of a variety of codecs (3000MXP, C40, Polycom VSX7000e) and a Tandberg Content Server.  All codecs have multisite (no MCU is being used) and TMS is used to schedule the calls.  The connections are internal H.323 via a Cisco VCS.

The section below is copied from the TMS call log, and shows the user repeadedly trying to send the presentation ("Duo Video opened"), which works for a few seconds before unexpectedly stopping ("Duo Video closed").  The "WCHRE Auditorium" is the 3000MXP endpoint:

11/09/2013 12:20 network service network service WCHRE Auditorium H323 Dial => 03839xxx1

11/09/2013 12:20 network service network service WCHRE Auditorium H323 Dial => 03839xxx0

11/09/2013 12:20 network service network service WCHRE Auditorium H323 Dial => OnDemand49A20386

11/09/2013 12:20 network service network service WCHRE Auditorium Call Connected: 03839xxx0

11/09/2013 12:20 network service network service WCHRE Auditorium Call Connected: OnDemand49A20386

11/09/2013 12:20 network service network service WCHRE Auditorium Call Connected: 03839xxx1

11/09/2013 12:28 network service network service WCHRE Auditorium Duo video opened. Source: WCHRE Auditorium

11/09/2013 12:29 network service network service WCHRE Auditorium Duo video closed. Source: WCHRE Auditorium

11/09/2013 12:30 network service network service WCHRE Auditorium Duo video opened. Source: WCHRE Auditorium

11/09/2013 12:31 network service network service WCHRE Auditorium Duo video closed. Source: WCHRE Auditorium

11/09/2013 12:31 network service network service WCHRE Auditorium Duo video opened. Source: WCHRE Auditorium

11/09/2013 12:31 network service network service WCHRE Auditorium Duo video closed. Source: WCHRE Auditorium

11/09/2013 12:32 network service network service WCHRE Auditorium Duo video opened. Source: WCHRE Auditorium

11/09/2013 12:33 network service network service WCHRE Auditorium Duo video closed. Source: WCHRE Auditorium

11/09/2013 12:33 network service network service WCHRE Auditorium Duo video opened. Source: WCHRE Auditorium

11/09/2013 12:34 network service network service WCHRE Auditorium Duo video closed. Source: WCHRE Auditorium

11/09/2013 12:35 network service network service WCHRE Auditorium Duo video opened. Source: WCHRE Auditorium

11/09/2013 12:36 network service network service WCHRE Auditorium Duo video closed. Source: WCHRE Auditorium

I would be very greatful if anyone can shed some light on this problem.

Regards,

Damien Jones.

[Martin Koch]

Check what kind of routers / firewalls you have in the signaling path from

endpoint > call controll (vcs-c?) > endpoint.2

It sounds like some layer3 component with additional intelligence, like a h323/sip aware firewall,

an alg or something else like that what messes with your signaling.

also if you have recently upgraded or added new systems with tc6.x, they increased the ports

for BFCP, so if you have strict port filters, check that tcp+udp/5070-5077 are openend up.

(see tc6 release notes or admin guide)

Please remember to rate helpful responses and identify helpful or correct answers.

[damienjones1]

Hi Martin, thanks for responding.

I've spoken to our network administrator and can confirm that there is no FW traversal taking place, all connections are on the same WAN.  The routers used are Catalyst 6500 and Nexus 7010 & 7009.  Assuming "alg" stands for application layer gateway, there is nothing like that running on the network according to the admin.

One of the endpoints is a C40 running TC6.0.1.65 (the others are 3000MXP and a Polycom VSX7000e) but as this is all on the same network, I don't think increased port numbers is the issue. 

Calls are routed via the Cisco VCS control (version X7.2.2), although I'm not quite sure how/if the h.329 / data presentation streams are routed via the VCS or not.

If you can think of any other avenues I can persue, please let me know.  I have increased the TMS logggin level from 1 to 2 to see if I can capture a bit more info on the issue, but given that it only occurs intermittently, it may take some time for this to yeild results.

Thanks again for any assistance or suggestions.

Regards,

Damien.

[Martin Koch]

Its never the worst idea to upgrade all systems to the latest software version.

Also try to identify if there is any kind of pattern.

If the problem also occurs with only cisco devices it is a very high chance that

it is some firewall somewhere. The signaling sticks to the vcs, the vcs is often in

a server network and the server networks have a  higher chance of having a

firewall.

Features which break something are often enabled by default and might not

even show up in the config unless disabled.

It would not be the first network which "definitly has no firewall" which at the end had one :-)

Please remember to rate helpful responses and identify helpful or correct answers.