The logical boundary of protection in Zerto is anything within a vCenter, as one ZVM is required per vCenter. If you vMotion a VM to another vCenter then it will break the protection group and the VM will no longer be protected.
You can however vMotion/Storage vMotion a VM between any datacenter, cluster or host within a vCenter without any interruption to the replication as long as a VRA is installed and running on the target host.
Your 2 options are therefore:
- Delete the protection group, but keep the target disks, vMotion the VM to the new vCenter and re-create the VPG with pre-seeding.
- Use Zerto to Move the VM to the target vCenter and reverse the replication, allowing you to move larger groups of VMs simultaneously than a long distance vMotion, but with the downside of a few minutes downtime.
- Register a host from the target site in the source vCenter, deploy a VRA in a separate VRA group, vMotion the VMs, perform whatever action the vMotion between vCenters was needed for then vMotion the VM back.
Which of the above would work best for you? Please can you share a bit more about your use case for vMotion between vCenters? Many thanks,
Hi Joshua, Thanks for the response. We are currently evaluating what makes the most sense for our vSphere environment. We currently have 3 datacenters and some remote offices each with their own vCenter instance. Looking at pulling them into one centralized vCenter instance to simplify management but also considering keeping our current design. We just implemented Cisco OTV which extends our layer 2 across our datacenters which brought up the cross vCenter vMotion question. We are currently on 5.5 U3 so we couldn’t test it out yet.
In this configuration I’d recommend having having each remote office registered to a central vCenter as a separate datacenter object.
You should then deploy VRAs to the hosts in each remote office in a separate VRA group, to indicate they are in a separate physical location.
With this configuration you can then create protection groups within the vCenter (enabled in advanced settings) replicating from the remote sites into the main DC for disaster recovery. You can then also leverage vMotion and Storage vMotion between the hosts in the remote site or from the remote site to the central DC without without breaking protection.
Any questions let me know. Thanks,
Travis – how about keeping the 3 “major site” VCs and just reducing the remote sites to a single VC? Or is that the plan and I’m just reading into it too much?
I don’t think we need to get into the design decisions of the vSphere environment as a whole since there are too many variables involved to hash it out via a forum.
From a Zerto standpoint it looks like a centralized vCenter is the way to go. Especially with vm’s that have to ability to move across datacenters.
Thanks for your responses!