Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

vcs expressway cluster

Hi,

We are planning to deploy VCS Expressway Cluster consisting of two expressway.  Both the Expressway will be enabled with Dual Interface. Need some clarity with respect to the ip and dns. following is the example setup.

eg

vcs expressway A: 192.168.10.2  asssigned to first interface.   192,168.20.2 assigned to second interface and nattted to 77.100.42.23

vcs expressway B: 192.168.10.3 assigned to first interface. 192.168.20.3 assigned to second iinterface and natted to 77.100.42.24

cluster name : video.exampledomain.com

FQDN:  video.exampledomain.com  will be  will be mapped to both the public ip address 77.100.42.23 and 77.100.42.24 with the required SRV records.

When movi client connect to video,exampledomain.com the traffic will be routed to 77.100.42.23 and 77.100.42.24  in round robin fashion.

Kindly confirm  is the above is correct way of deploying and also what will happen if vcsexpressway A is down. always the first call will fail because the rqurest will be sent to 77.100.42.23  and Second call attempt will succeed.

Krishna.

1 ACCEPTED SOLUTION

Accepted Solutions
Gold

Re: vcs expressway cluster

Hi Krishna,

your deployment plan looks good, assuming that 192.168.10.x and 192.168.20.x are different subnets. Please make sure to set up clustering on LAN1 (The non-NATed interface) which seems to be part of your plan.

Regarding the call behaviour when one of the VCS's is down, assuming that your SRV records have been properly set up and that the remote calling party has proper DNS SRV support for its H323 and/or SIP implementation, the calling party should fail over to the working VCS if the initial connection to the downed VCS fails. In any case the outcome will largely depend on the behaviour of the calling party and its capability to switch over to the other VCS when its attempt to connect through the downed VCS fails.

Hope this helps,

Andreas

4 REPLIES
Gold

Re: vcs expressway cluster

Hi Krishna,

your deployment plan looks good, assuming that 192.168.10.x and 192.168.20.x are different subnets. Please make sure to set up clustering on LAN1 (The non-NATed interface) which seems to be part of your plan.

Regarding the call behaviour when one of the VCS's is down, assuming that your SRV records have been properly set up and that the remote calling party has proper DNS SRV support for its H323 and/or SIP implementation, the calling party should fail over to the working VCS if the initial connection to the downed VCS fails. In any case the outcome will largely depend on the behaviour of the calling party and its capability to switch over to the other VCS when its attempt to connect through the downed VCS fails.

Hope this helps,

Andreas

New Member

Re: vcs expressway cluster

Hi Andreas,

Thanks for you reply. Is the switchover transparent to the user. for example if somebody wants to dial conferencroom@exampledomain.com will the first call attempt succeed eventhough the primary vcs is down?

Krishna.

Gold

Re: vcs expressway cluster

Krishna,

the switchover behaviour depends on the capabilities of the calling party, with regards to DNS SRV.

If you have DNS SRV records set up for H323 and SIP for the exampledomain.com domain, pointing to each VCS-E with equal priority and weight, the calling party should attempt to set up a connection with either of the peers randomly. If the calling party happens to attempt to connect to the downed VCS and not suceeding with this, ideally and normally it should then give up and make a new attempt, this time connecting to the other,working peer as specified in the SRV records, and this will then succeed. This process would be transparent to the calling and called party, but again this depends on the SIP/H323 implementation of the calling device (Or the gatekeeper/proxy which initiates the connection to your VCS-E's).

Regards

Andreas

vcs expressway cluster

Yes, the call is going to success. You will notice that usually the video endpoint register h323 with one VCS and SIP with with the other VCS so the video enpoints will be always available if you are running both H323 and SIP.

Also you can

Set the H.323 Time to live to 60 (seconds) so that if a VCS becomes inaccessible

to an endpoint, the endpoint will re-register quickly with another peer (VCS configuration > Protocols > H.323).

1967
Views
0
Helpful
4
Replies