Do we need a VR server at the protected site ?

[vylco]

Hi there,

I wonder if i need to have the same number of VR server at the both site (protected and recovery) or just at the recovery site or SRM use the VR server  of the recovery site for the failback ?

Regards.

[mvalkanov]

Hi,

The VR server accepts the incoming replication traffic from the source ESXi and should be as close as possible to the hosts that have access to the target datastore(s).

If you use VR with SRM, then you definitely have separate vCenter servers for the source site and target site. A VR server appliance can belong to only once vCenter inventory and has access only to the hosts and datastores in that vCenter inventory. VR w/o SRM supports also a single vCenter inventory for source site and target site.

With VR 5.1 and next, a VR server is present in the combined VR appliance (VRMS + embedded DB for VRMS + VR server). You would need to deploy additional VR server only appliances, only for load balancing and/or better access to the target datastores.

Regards,

Martin

[basher · Accepted solution]

Hi,

Having VR servers at the protected site is indeed mandatory for failback. I would recommend having the same number of servers or you may hit some timeouts once you reverse replication (first step of failback).

Best regards

Stefan

[vylco]

Hi, thanks for these good answers.

Best regards.