We need to upgrade operative system in cluster nodes and AGs from RHEL 7.4 to RHEL 7.9 and then, from RHEL 7.9 to RHEL 8.6 in a Managed tenant.
We have three nodes.
We know to do this we have to stop Dynatrace services in the machines, do the o.s. upgrade and then start Dynatrace services. We've seen two procedures for that:
1.- Stop/start the nodes one by one: https://www.dynatrace.com/support/help/managed-cluster/operation/start-stop-restart-node
2.- Stop/start the whole cluster: https://www.dynatrace.com/support/help/managed-cluster/operation/start-stop-restart-cluster
We would like Dynatrace was working during the upgrade. So we think about doing the upgrade node by node. But we are worry about split brain because there will two nodes running till the other one is being upgrade.
If we stop/start the whole cluster, we prevent split brain issue. But we didn´t have monitoring.
So, what are the recommended actions in this case?
Thanks in advance.
Solved! Go to Solution.
I've been through the same situation with a 3 nodes cluster, don't worry about split-brain as the upgrade of each node should not take much time, of course, it will be based on your upgrade plan, so try to make sure that you are going to upgrade all cluster nodes at the same down time window and you can go with upgrade OS on the cluster nodes one by one.
We just updated the operative systems sucessfully following this link:
Thanks a lot for your comments.