We have a cluster in production with 5 members in RHEL7 and we want to perform a rolling upgrade to RHEL8. The idea would be to upgrade the nodes to RHEL-8 one by one and have a cluster with members on rhel7 and rhel8 working perfectly until the last node is upgraded.
Are there any known incompatibilities that make this not possible? Of course, the cluster has primary and replica shards and stopping one node does not cause incompatibilties in any index.
I’ve seen some issues with regards to performance while upgrading between RHEL versions. Not any incompatibilities, but since you change things like kernel version, it’s best to do some testing to make sure things work and perform as expected.
I know this is a rather generic answer. Specifically, there shouldn’t be any incompatibilities, but depending on how critical your cluster is, you’ll probably want to check first. Or at least upgrade one node and let it run for a while until you move to the others.