Disaster Recovery

 View Only
  • 1.  HA and DR together

    Posted Mar 20, 2025 10:03 AM

    We have Live Site Recovery deployed and operational for a single VM over two geo-separated sites, physically connected over Dark fiber of 5 miles distance.
    We use VM based VR as each of two hosts uses its internal disks as storage. 
    Storage usage of the VM is 1.5 TB. DCI is 1 Gbps with 2 msec latency. 
    Now we consider to deploy HA for the same VM at Site A only while we want to keep the Live Site Recovery in place. 
    So, we consider to move the VM from the internal storage to a shared storage reachable from both sites via iSCSI. 
    Next we should add Host B to Cluster A and enable HA. 
    We assume Host B can be a part of both Custer A (for HA) and Cluster B (for DR), and we don't need to change its Mgmt IP address.
    Please confirm if the above is feasible. 



  • 2.  RE: HA and DR together

    Broadcom Employee
    Posted Mar 20, 2025 10:07 AM

    That's not how it works, a host can only be part of 1 cluster. So you cannot add Host B to Cluster A and still have it actively participate in Cluster B.




  • 3.  RE: HA and DR together

    Posted Mar 21, 2025 10:25 AM

    Thanks for the clarification!

    We consider HA just to have minimum data loss when the VM is moved from Host A to Host B as well as automatic switchover. The deployed Liv e Site Recovery supports around 5 minutes data loss with RPO of 5 minutes as well as it manual switchover only.

    So, it seems now we may have only following options, i suppose:

    1) Use Veeam or NAKIVO as they support automatic switchover and almost no data loss. With this option we don't need the shared storage and any extra hardware (server).

    2) Add one more server to Cluster A, move the VM from internal storage to the shared storage at Site A, reconfigure Live Site Recovery.

    Obviously, the option 1) is more efficient as it doesn't require huge expenses.