I am trying to envision what this would look like in a Nutanix environment for planning purposes.
Currently we have two datacenters. We have one large vcenter cluster (Cisco UCSs, Synchronous Nimbles, Nexus vpc pair, L2 stretch on our own dark fiber (multiple 10gb links)) stretched across both DCs. This allows us to vmotion all VMs from one side to another, take down half our hosts, or even a whole DC without any issues.
Thinking what this would look like if we were full Nutanix and AHV I am pretty sure this would not work (1 large cluster as we couldnt take down 1 DC since that would take out half of the hosts\storage). Correct?
If so, I am thinking this would require at least a cluster at each DC, and then you metro-availability to uses the real-time sync. Does that sound about right from a high level?
There are two points, first is stretched cluster (one cluster span two physical sites), second is Metro availability (One cluster replicated with 0 RTO to another cluster.
First one needs advanced networking and apply RF2 or RF3 (RF2 means one node or one disk failure) (RF3 means two nodes or disks failure) tolerance. Then below is not correct, it depend on RF and available resources for self-healing
Second one needs two pair of clusters one replicate to another with 0 RTO, and you can make failover planned and unplanned manual or automatic using witness VM. It requires <5ms network latency between the two clusters.
Yes with full nutanix and ahv you need two clusters. 1 in each dc. You can configure synchronious replication to keep this in sync and do the same as with you vmware metro availability.
i’ve created a blog post about this: https://next.nutanix.com/community-blog-154/convert-your-cluster-to-ahv-and-configure-leap-for-disaster-recovery-40970