VCS Clustering

[reji.ks87]

Hi All,

I have a question on VCS clustering. Can we cluster two VCS-Control which has a Traversal zone to a single VCS-Expressway.

Basically two VCS-C clustered and a single VCS-Expressway.

Thanks, Best Regards, Reji.

[Jens Didriksen · Accepted solution]

Yes you can.

Ensure the non-master does not have any settings configured apart from the default zones and links as it will take its configuration from the Master, including traversal zone settings.

Suggest you take a look at the cluster creation and maintenance guide relevant to the software version your systems are running; http://www.cisco.com/c/en/us/support/unified-communications/telepresence-video-communication-server-vcs/products-installation-and-configuration-guides-list.html

/jens
Please rate replies and mark question(s) as "answered" if applicable.

[reji.ks87]

Hi Jens,

I have one more question regarding DNS SRV record. I have the following.

Domain: example.com

Master VCS FQDN: vcs1.example.com

Peer VCS FQDN: vcs2.example.com

Cluster FQDN: vcs.example.com

How / where can we create a DNS SRV record.

If its under example.com then how an endpoint with GK address as vcs.example.com register to the cluster. 

Or can we create the SRV record under FQDN vcs.example.com.

As per our server team they can only create under domain example.com.

Thanks for your reply.

[Patrick Sparkman]

What we did was have each of the VCS cluster peer A records point to the cluster FQDN, and register the endpoints to the FQDN. 

SRV records would be configured under the main domain for inbound calling into the VCS cluster for external endpoints. 

[Jens Didriksen · Accepted solution]

We do it a bit differently; we do create a separate FQDN record for both master and peer which includes reverse look-up, then we also create a FQDN record for the cluster which points to both IP addresses of the master and peer, but without reverse-lookup.

So we have vcs1.example.com and vcs2.example.com which resolves to their individual IP addresses and vcscluster.example.com which resolves to the two VCS IP addresses.

SRV records points to both master and peer; (the values, i.e. ttl, weight etc  are just examples and should be adjusted to suit your deployment)

_sips._tcp.example.com. 43200 IN SRV 1 1 5061 vcs1.example.com
_sips._tcp.example.com. 43200 IN SRV 1 1 5061 vcs2.example.com
_sip._tcp.example.com. 43200 IN SRV 1 1 5060 vcs1.example.com
_sip._tcp.example.com. 43200 IN SRV 1 1 5060 vcs2.example.com

_h323ls._udp.example.com. 43200 IN SRV 1 1 1719 vcs1.example.com
_h323ls._udp.example.com. 43200 IN SRV 1 1 1719 vcs2.example.com
_h323cs._tcp.example.com. 43200 IN SRV 1 1 1720 vcs1.example.com
_h323cs._tcp.example.com. 43200 IN SRV 1 1 1720 vcs2.example.com

_h323rs._udp.example.com 43200 IN SRV 1 0 1719 vcsc1.example.com

_h323rs._udp.example.com 43200 IN SRV 1 0 1719 vcsc2.example.com

In our older MXP system we set the h.323 gatekeeper and SIP proxy addresses to use the cluster FQDN, i.e. vcscluster.example.com.

In our C- and SX-systems, we also set the h.323 gatekeeper address to the cluster address, but these systems allow for more than one SIP proxy address to be specified, so we specify both master and peer addresses in these.

You could just use the cluster address, but we've found the above works best for us.

Just a word of warning: to use the above, you must have split DNS as you do not want these records to be seen by external systems, just as you don't want the SRV records pointing to your VCS-E to be seen by your internal systems.

If you don't have split DNS, then your best option is to not use SRV records at all for your internal systems. In this case you would set Gatekeeper discovery to manual and register using the cluster address only.

/jens

Please rate replies and mark question(s) as "answered" if applicable.

[reji.ks87]

Hi,

We did the Clustering and finally this is the only Alarm. "Expected default link between Default subzone and the Cluster subzone is missing". Thru CLI tried "Xcommand DefaultLinksAdd"

but till the same. Any solutions.

Thx.

[Jens Didriksen]

Haven't encountered that one before, don't know if this can be of some help; https://supportforums.cisco.com/discussion/11427986/vcs-control-configuration-warning - from awinter2 reply onwards.

/jens

Please rate replies and mark question(s) as "answered" if applicable.

[reji.ks87]

After clustering found that "Expected default link between Default subzone and the Cluster subzone is missing" Tried thru serial log in "Xcommand DefaultLinksAdd"  but no luck.

While checking Links betweeen defaultsubzone and clustersubzone node 2 is blank. Deleted this link and re entered "Xcommand DefaultLinksAdd". Done Alarms gone  Same Link automaticaly created with node 1 as defaultsubzone and node 2 as clustersubzone.

thanks for all the support.