Some DataMiner deployments span across different datacenters and geographical locations. In these cases, it would make sense to install the Primary and Backup DMAs in separate datacenters to achieve better redundancy e.g. if one entire DataMiner goes down the backup agent in the other datacenter can continue operating normally.
That makes sense in theory, but in practical terms is it recommended or even possible to use such failover architecture? If so, what are the communication requirements between the Primary and Backup DMA (physically hosted in distance locations)?
I believe Primary and Backup DMA must be part of the same subnet. If that is truly the case then hosting them in separate datacenters might not be feasible since most of the times there are firewalls and routers between datacenters.
Please share some practical examples on how geo-redundancy can be achieved in the context of DataMiner failover deployments.
Miguel, we indeed encounter the problem with the virtual IP we use in our failover setups. We have scheduled a work-item on the list to provide a solution for this subnet problem. Do note that the first use-case we'll be evaluating is linked to a AWS deployment of a failover system. We aim to have this setup validated by DataMiner 10.1.1 (Mid January 2021)