UCCX Reboot

[pcromwell]

I was told that if you have UCCX with High availability, then anytime a reboot of a node is required, you need to disable High Availability and then when the Node comes back you would re-enable High Availability. I am about to do an upgrade and am at a loss of how to disable HA. Am I missing something obvious?

[Deepak Mehta]

I had HA deployment and i rebooted the cluster one by one.

Starting with Primary Master node and wait until this comes up fully then went on to reboot the secondary node.

I believe once you reboot the HA is automatically broken and it works in island mode .

[Deepak Rawat · Accepted solution]

First of all, you will need to understand that UCCX works on the concept of Active-Standby. What this means is that at a given point of time only one server will do everything related to agents login, call routing, script execution etc. Now coming to your question how to disable HA. Once you will restart the primary UCCX that will by default break the HA and mastership will fail over to secondary server. Once it comes back the HA will be restored again however when you will restart secondary server HA will break again and mastership will fail over back to primary. However once secondary server comes back after the restart HA will be restored again and master ship will remain with primary UCCX server only.

Hope the above is clear hence simply perform the reboot I.e. first primary once it comes up then go for the restart of secondary server.

@Deepak M: Your understanding about HA being broken on reboot of any server in cluster is correct however it will go in Island mode is not correct. Island mode only happens when the secondary UCCX server is over a WAN link and the link goes down. In that scenario only both servers think themelves as Master and both can perform activities like calls routing, agents login etc simultaneously. However once the WAN link comes up again one of the server will drop the mastership and then again work as normal I.e., one of them will be Master and other will be secondary.

Refer to SRND for more details.

Regards

Deepak

[Deepak Mehta]

Thank you for correction. I used a wrong word. 

[pcromwell]

Thanks both Deepaks for the responses, I upgraded primary into inactive partition without switch version, I then upgraded secondary without switch version. Then did the Primary reboot and waited till all services running. Then the Secondary and all worked well. The only odd thing after that, was although I was expecting CAD to ask to be upgraded, It needed the CAD config tool to be run first as it didn't seem to know where to get the CAD file from, I then needed to upgrade Java. bit painful just for a 9.0.2 SU1 to SU3

[Deepak Rawat]

Running Client Configuration Tool is also a mandate after performing the upgrade and only then CAD/CSD can be upgraded. Hence what you experienced there is again something a default behaviour and not anything erroneous. Anyways the upgrade activity had been done successfully, clients had been upgraded hence I think it is time to sit back and relax :)

Regards

Deepak