Greetings,
Can anybody comment on this upgrade plan?
We have 4 nodes:
1 - Primary Admin/Secondary MnT
2 - PSN only
3 - PSN only
4 - Primary MnT/Secondary Admin
We know the upgrade order will be 4 - 3 - 2 - 1.
We are testing this in the lab and will perform all appropriate precautionary measures. However we are still concerned that an unforeseen problem will crop up in production after the upgrade and want the ability to revert back quickly if needed.
After upgrading nodes 4 and 3 we will have one Admin/Mnt Node and one PSN in the new deployment, but still have one Admin/Mnt Node and one PSN in the old deployment.
All NADs are set to use either node 3 or 2 as primary radius server with failover to the other PSN. If after upgrading nodes 4 and 3, we stop ISE services on node 2, all NADs will be using node 3, which is in the new deployment. If we experience problems, we can bring up services on node 2, stop services on node 3, thus forcing all NADs to use the old deployment. If after one week we are satisfied the upgrade has gone well, we can upgrade nodes 2 and 1 to complete the process, NADs will go back to load balancing with failover across the 2 nodes.
Will this work? What issues might we encounter? Thanks.
