We are planning to build the DMA on virtual environment. We used the Node Calculator to get the estimated resources required on our project.
May someone advise if my understanding below is correct.
- For DMA high-availability we need 4x DMA node + 4x DMA node is this correct?
- For Cassandra and Elastic Search, do we need also to duplicate the VM similar to above DMA? or we can have 3x Cassandra VM and 4x Elasticsearch VM then from the Hypervisor we can provision the redundancy.
To have a DMA failover (one DMA team up with a backup server, which is constantly in sync, so it's ready to take over if the online DMA fails), you always need a pair of DMAs, so if your DataMiner cluster is designed to have 4 running DMA, you will need 4+4 DMAs to team up in a failover pair.
About Cassandra and Elasticsearch, if you configure this as separated clusters from DataMiner, you only need at minimum 3 nodes on each cluster to provide you redundancy, in case one of them fails.
I recommend you this docs page where the supported architecture for storage are described: Supported system data storage architectures | DataMiner Docs