17 Feb 2021 05:21 AM - last edited on 30 Aug 2022 03:22 AM by MaciejNeumann
We're planning to move our POC cluster to a new Producton cluster .
2 POC nodes are installed on a virtualized environment and the new production nodes will be 4 physical machines with better CPU/RAM/Storage configuration.
According to the Dynatrace documentation (1, 2, 3) new nodes "must" be in identical configuration as the old ones, so if we go with the data replication method by adding new cluster nodes, we will have incompatibilities.
How do you think we should proceed?
Solved! Go to Solution.
Yes, all nodes should be equal to ensure nodes are equally balanced. In case of upgrade or failures, other nodes can equally handle the load.
In case this is a temporary solution for the migration time, it's fine they are not the same (with the side-effect I mentioned).
Good luck with migration!
That must is for supported production operation, not a strict technical restriction.
For a short term you will be fine if the new machines have at least the same size / are bigger and you have the same number or more nodes in the new cluster.
So for you migration use case with the replication method you will be fine.
Hello @MUSTAFA SERHAT B.
In my opinion, the cluster backup and restore procedures should be suitable in your situation.
Secondly, you are also going to provide better hardware and more nodes. According to the below-documented information.
It depends on how important the transaction detail data is. With the Backup method you lose the transaction detail data for the last 10 days as it is not backed up.
If that is ok you can use the backup, if not stick with the replication method where the data will still be available on the old nodes till phased out.
Thank you guys for all the suggestions. The idea of starting from scratch and reconfiguring all the months of work was making me nervous.