UCCX Hardware Upgrade

[dvanzee]

Hello All,

Currently I have a UCCX cluster(7.0.1sr5) which is running on old 7845 -H1 hardware. I have  been tasked with upgrading the hardware.

1- Does anyone see an issue with using two types of server hardware? I have a 7845-H2 & 7835-H2. Will this hardware work to replace the current 7845-H1 hardware?  Will Cisco support it?

2-  Can I simply remove the "salve" node, build my first UCCX server and add it to the existing cluster? Make the new hardware the master server retire the old, build my second UCCX server and add it in?  Is that possible, is it an acceptable upgrade path?

Any information or insight would be much appreicated.

[Robert W. Rogier]

dvanzee,

     Well, that's not techinally supported.  Both machines need to be equivalent class machines.  They both need to be 7845 or both 7835 to survive.  Having them  unequal like you are suggesting puts a difference in the actual performance of the machines.  If you read the SRND you will see that a 7835 can not support as many people, ports, etc as a 7845.  The key thing is that both machines need to have the same throughput, i.e. same processors, same RAM, equivalent drives, etc.  If the only real difference is that one is HP and one is IBM, that's ok.  You just can't run a 7845 and an 7835 together.  I would think you might be able to do it if you use VMWare and each machine is on the compatiblity matrix.  Then you could allocate equivalent resources on each side so that one machine isn't overpowering the other.  You'll end up with extra processor cycles on the 7845, but that would be far better than having a 7845 that can support 300 ports try to do a failover to a 7835 that only allows 150 ports.  Things would go drastically wrong in that case.  Hope this helps.  Let me know if you have further questions.

Regards,

Robert W. Rogier

CSE -- TAC, UCC

[frzhang]

There is no specific requirement to have equivalent server in HA cluster. We used to have this type of problem in UCCX 7 cluster. But to my understanding, you server is supported as long as it's listed in the compatibility information for that release. If running different server in the cluster, you have to understand the capability of the server won't be the same. So, if you are approaching the limit of the server capacity, the failover server with less capability will not handle all the work as the publisher. I think this should be easy to understand.