Hi Ram,

Not sure if a Failover is the best answer for you atm. If you want to have the ability to be fully operational when your main site is completely lost (also the equipment on the main site). Would it then not make more sense to have two separate DataMiner (DM) systems?

The main DM system would monitor and control everything on your main site. The DR DM system would then monitor everything only from the DR site. If you need to have some of the equipment in your main DM system of the DR site, you could replicate the element from the DR DM system to your main system. The advantage of this over a failover system would be that it would be more robust, especially for network limitations between both sites. The downside is that you have two systems to manage and you might need to put other safety in to cope with servers going down (e.g. through VM snapshots or through backups).

Assuming that the sites are Izegem and Brussel, this is how the minimum (depending on the scale of your sites) FO setup would look like:

It is advised to run every block on a dedicated VM, so here you would need 3 VMs on both sites. With this, you have two separate Elastic Clusters (EC). DM will offload data in files when one of the two is unavailable to avoid data loss and push it back once the EC is back available, but there is no active syncing between the two ECs. So your data could be out of sync on EC level with this setup. At any point in time, only one of the two DMA's for every FO pair is active. Both agents in the FO pair will be like a copy of each other (creating an element on the active DMA means it will be on the FO as well). The active DMA(s) will be handling everything and in case of failure the FO can start up, so you still have the startup time of your FO (roughly 15 min for a fully loaded DMA). So all polling is done by the active DMA and might need to go over the connection between your sites. If you start to scale it up, it might look like this:

This is how the minimum setup will look like for two separate systems:

Here it would mean again on both sites 3 VMs for the minimum setup, but every system would be communicating with the equipment on that site. If you need the elements to be available on the other side you can use replication between the two sites. This means only communication between DM systems for replication and no need for DM to reach the equipment on the other site to get visibility (less firewall/routing work). With this minimum setup if you have no redundancy through DM. You could rely on VM snapshots or backup restore mechanisms to get the system back after a server loss or you could also have redundancy on DM level on both sites:

Hope this makes things a bit clearer on what options you have. Depending on the scale of your system and what is most important for you one of the two might be better suited. Let us know if something is still unclear for you.

More information can be found on Supported system data storage architectures | DataMiner Docs and on DataMiner Compute Requirements - DataMiner Dojo