"Metro" is actually a vSphere concept from the beginning.
Hyper-V has syncronous replication, but not a single Hyper-V failover cluster spanned between sites
AHV does not have sync replication, though we do have that on the roadmap and it is under active development.
It is mean that Disaster recovery in Nutanix Platform:
vSphere support: Metro Availability,Syncronous replication and Async replication
Hyper-V support Syncronous replication and Async replication
and AHV only support Async replication
Is it right and do we have any link to check that which feature work in which hypervisor platform?
@chenzh4 based on what I've known, Metro Availability and Synchronous DR are the same at one in operation.
Basically, Threre are just only 2 types of DR as you might know, Sync DR and Async DR. In the case of ESXi hypervisor, Sync DR is renamed or called by other name (Metro Availability).
I checked some documents and blog, and found that Metro availability be set in 1 vmware vsphere cluster, but sync replication DR should be set in 2 vmware vsphere clusters.
but seldom mentioned that AHV DR whether Sync DR or Async DR. Does AHV suypport Sync DR or Async DR?
AHV today supports only Async DR. Sync is on the roadmap but not currently supported.
The difference between Metro and Sync replication is indeed whether you are using a single vsphere HA cluster (which spans the two Nutanix clusters), or separate HA clusters.
When you're using separate HA clusters we refer to this as Sync rep. Nutanix supports Sync rep with both ESXi and Hyper-V (where you'd have separate Windows Server Failover Clusters). True Metro is only supported on ESXi.
Whether you're doing Sync rep or Metro, both are configured at the Nutanix level essentially the same way, by setting up a Metro Availability protection domain. In the case of Metro you can utilize the witness, in the case of sync rep the witness is not supported.
In the case of Metro, ESXi HA mechanisms manage VM movement and restart between sites. In the case of sync rep, VM's must be moved manually (or scripted) by unregistering and registering in the appropriate clusters.
Hope this helps clarify.
It is very clear for me. Thank you very much.
I read a lot of release notes and did not find the async DR in supportability of Features, only found synchronous replication and MA features.
Now I got your clarification. Suggest Nutnaix can maintain a detailed supportability of Features by Hypervisor.