VCS upgrade to X8.1 from X.7.1 no video passed

Unanswered Question
May 8th, 2014
User Badges:

I upgraded from X.7.1.1 to X.8.1.1. I no longer get video or audio across traversal.  everything works internal.  Video/Audio also works if registered to the expressway calling outside of network.  It is only traversal link.  I have asked them to confirm the ports 50000 and 50001 are opened.  I have also now asked them to open the full port range 36000 – 59999 UDP .  I have tried changing the port range for traversal media.  to full range.  I did not leave it this way.  I would have had to change the Ephemeral port range.  So it is back to the media port range is back to 50000 - 54999 UDP.  Any suggestions?

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Chad Patterson Thu, 05/08/2014 - 17:25
User Badges:
  • Cisco Employee,

Hey,

My suggestion here would be to take a packet capture and diagnostic log from both the control and the expressway while placing a call that will go across the traversal zone. First look in the diagnostic log from either the control or expressway and look at the invite message and 200 OK message from the calling and called endpoint. In the SDP you will be able to see what port was set for Video and what port was set for audio. Specifically look for the m=video and m=audio lines in the Invite and 200ok.

 m=audio 51780 RTP/AVP 108 104 105 9 18 8 0 101

 m=video 51810 RTP/AVP 97 126 98 34

These lines were pulled out of an invite. Here you can see the endpoint who sent this invite message requested for audio to be sent to port 51780, and for video to be sent to port 51810.

Next open the packet captures in wireshark and go to Telephony > RTP > Show All RTP Streams. Note the source/destination ports being used from the endpoint to the VCSc, from the VCSc to VCSe, from the VCSe to the far end and vice versa. Are the ports being used in the packet capture the same ports that were requested in the signaling (Invite and 200Ok)?

Make sure that what was requested what actually what was being used and that nothing in between the endpoints is modifying the RTP source and destination port being used.

Also, were both the control and expressway upgraded from 7.1.1 to 8.1.1?

chrischelvy Thu, 05/08/2014 - 17:28
User Badges:

Hi,

         if you have a NAT enabled on the X7.1 , then please make sure that NAT settings are in place after the upgrade.

 

Thanks

Chris

Martin Koch Thu, 05/08/2014 - 18:08
User Badges:
  • Red, 2250 points or more

It might be handy to know as well what call protocol and endpoints were used, if other mechanisms like STUN/TURN/ICE are used as well,

if all calls fail like that or if there are some which work or behave different, how the network in between all components looks like,especially

if NAT, Firewalls or ALGs are present, ...

(remember to check that there are no layer-3 sip/h323 "features" present .

 

Are your VCS-E and -C both on X8.1.1 or do they run on different software versions? If so which? X7 and X8

even if said in the earlier release notes should work together. But anyhow I would recommend that you have X8.1.1 on both.

 

I also like to use a network capture. If you know how to use tcpdump it can be very helpful as you find it on the VCS,

on C/ES/MX/SX endpoints (or like said using wireshark on a laptop).

 

 

Chris Swinney Fri, 05/09/2014 - 05:28
User Badges:
  • Silver, 250 points or more

I will also stick more oar in here, in reference to what Martin said, have you also upgraded the VCS-C, and is there any kind of firewall between your VCS-C and endpoints?

If you are successfully making internal calls and devices are registering with the VCS-C, chances are that media is flowing directly between the endpoints and NOT via the VCS-C (am am assuming a few things here). Of course if you have updates the VCS-C, then the same things apply to that as you have been aware of with regard to the traversal zone to the VCS-E.

 

Cheers

Chris

Actions

This Discussion