12-26-2012 08:41 AM - edited 03-18-2019 12:20 AM
I would like to get some feedback to add a second VCS to my video infrastructure and clustering
Active VCS has the following
Solved! Go to Solution.
12-26-2012 11:50 AM
Cluster name and cluster pre-shared key should be the same on all peers in the cluster and local host name / system name should be unique on each peer.
Round trip time between each peer in the cluster should be under 30ms
Thanks,
Guy
Sent from Cisco Technical Support iPhone App
12-26-2012 03:52 PM
While VCS is running X7.1, for clustering VCSs…,
Configure same cluster name on all VCSs
Configure same pre-shared key for clustering communication on all VCSs
Make sure all VCSs has same feature option keys (i.e. provisioning, findme, H.323-SIP Gateway, etc.)
Configure appropriate DNS and NTP server address on VCSs (this parameter doesn’t have to be same within cluster VCSs)
And important (repeat what Guy mention as this is really important for clustering VCS) the round trip between VCSs within cluster VCS must be within 30 ms.
12-26-2012 04:27 PM
In addition on what Guy and Tomonori already said.
Also check that the vcss have exactly the same version number. Call licenses do not have to be the same btw.
Check what you have to do regards provisioning and tms if you use them as wel.
The best would be to read through the VCS Cluster and the admin guide you can find on the cisco site:
http://www.cisco.com/en/US/products/ps11337/tsd_products_support_series_home.html
As you did not say much about your deployment here there might be some additional or different taks be needed.
Consider reaching out to a Cisco TelePresence Partner / Consultant to help you.
As you write "Master", the cluster has to be seen as one, so you do not have a real master and slave
or higher or lower priority (just for the configuration you have the master role).
There are some hacks with pointing different priorities on the SRV records to the VCS but you
might end up in having trouble regard traffic flow of traversal or neigbor zones if you would expect for
example the NY one to be the "Master" which should handle all the traffic.
But you could handle a more distributed scenario with srv records and clever search rules for failover
even without a cluster, but then you might loose some benefits of the cluster (redundant provisioning, ...)
You can name them like this (that its called vcsc is just to make it better visible, but sure you could also just call it vcs, and sure domain.com should reflect your setup :-)
On both the same:
Cluster name: vcsc.domain.com
(FQDN for Provisioning, if you use findme this has also to be the internal sip server address)):
Domain Name: domain.com
NY:
System Name: vcsc-ny.domain.com
Local Host Name: vcsc-ny
NJ:
System Name: vcsc-nj.domain.com
Local Host Name: vcsc-nj
Then you generate A (PTR records are also handy) DNS entries as well as SRV records.
Please be aware that you should run the latest versions on your endpoints to properly support
SRV records, also 3rd party endpoint might have issues (but you could add additional SRV records for specific types).
Please remember to rate helpful responses and identify
12-26-2012 11:50 AM
Cluster name and cluster pre-shared key should be the same on all peers in the cluster and local host name / system name should be unique on each peer.
Round trip time between each peer in the cluster should be under 30ms
Thanks,
Guy
Sent from Cisco Technical Support iPhone App
12-26-2012 03:52 PM
While VCS is running X7.1, for clustering VCSs…,
Configure same cluster name on all VCSs
Configure same pre-shared key for clustering communication on all VCSs
Make sure all VCSs has same feature option keys (i.e. provisioning, findme, H.323-SIP Gateway, etc.)
Configure appropriate DNS and NTP server address on VCSs (this parameter doesn’t have to be same within cluster VCSs)
And important (repeat what Guy mention as this is really important for clustering VCS) the round trip between VCSs within cluster VCS must be within 30 ms.
12-26-2012 04:27 PM
In addition on what Guy and Tomonori already said.
Also check that the vcss have exactly the same version number. Call licenses do not have to be the same btw.
Check what you have to do regards provisioning and tms if you use them as wel.
The best would be to read through the VCS Cluster and the admin guide you can find on the cisco site:
http://www.cisco.com/en/US/products/ps11337/tsd_products_support_series_home.html
As you did not say much about your deployment here there might be some additional or different taks be needed.
Consider reaching out to a Cisco TelePresence Partner / Consultant to help you.
As you write "Master", the cluster has to be seen as one, so you do not have a real master and slave
or higher or lower priority (just for the configuration you have the master role).
There are some hacks with pointing different priorities on the SRV records to the VCS but you
might end up in having trouble regard traffic flow of traversal or neigbor zones if you would expect for
example the NY one to be the "Master" which should handle all the traffic.
But you could handle a more distributed scenario with srv records and clever search rules for failover
even without a cluster, but then you might loose some benefits of the cluster (redundant provisioning, ...)
You can name them like this (that its called vcsc is just to make it better visible, but sure you could also just call it vcs, and sure domain.com should reflect your setup :-)
On both the same:
Cluster name: vcsc.domain.com
(FQDN for Provisioning, if you use findme this has also to be the internal sip server address)):
Domain Name: domain.com
NY:
System Name: vcsc-ny.domain.com
Local Host Name: vcsc-ny
NJ:
System Name: vcsc-nj.domain.com
Local Host Name: vcsc-nj
Then you generate A (PTR records are also handy) DNS entries as well as SRV records.
Please be aware that you should run the latest versions on your endpoints to properly support
SRV records, also 3rd party endpoint might have issues (but you could add additional SRV records for specific types).
Please remember to rate helpful responses and identify
12-27-2012 11:24 AM
Thank you for taking the time to respond to my questions. I had a difficult time interpreting the CCO document that stated the Call Licenses can be different however all other keys must be identical. Since it didn't state the "other" keys is the Pre-Shared key lead to some confusion.
Now I'm trying to implement Option 1 Appendex 5 for DNS SRV. My C40 endpoint under H.323 has a field to enter the IP address of my VCS however it doesn't appear to accept a DNS name such as my VCS Cluster Name.
I have the same issue with my Polycom HDX 8006 where it has a field for the VCS IP address and no field to enter the VCS Cluster name.
12-27-2012 02:51 PM
Do you have H323rs SRV service record for domain you are configuring on C40 Endpoint as H323 Gatekeeper address?
For H323 Gatekeeper registration by using SRV, you must have H323rs SRV record.
H323rs service define registration service, entity supporting H.225.0 RRQ.
You may add this SRV on VCS cluster name and configure it as H323 Gatekeeper address on Endpoint.
If you haven’t configure this a part of SRV record, please add it and try it again.
H323rs service should configure with following parameters.
================
Service: H323rs
Protocol: UDP
Port: 1719
================
12-27-2012 02:54 PM
BTW, I have posted DNS SRV document a while ago, https://supportforums.cisco.com/docs/DOC-25395, which might become handy for you for configuring SRV service on your domain.
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: