vCenter DR

[pwilcox]

Customer has Production site A and  DR site B. Latency between sites is greater than 10ms. Customer plans to use SRM to protect management VM's on Site A.

Customer has 100+ remote field sites that are managed by site A vCenter.

In the event of disaster at site A,  SRM will failover management VM's including vROPS to site B. We are assuming DNS is set up so that change in IP address of vROPS will not affect it's ability to continue to monitor 100+ remote field sites.

Customer wants to know if there is a way that they can manage hosts's and VM's from site B after failover?

Here is what I have looked at:

• vCenter High Availability is out because of greater than 10 ms. Also not sure how that would work with SRM. Could not find any documentation.
• We can't do restore of vCenter from site A to B because that not work with SRM.
• They can manually register the hosts in site B's vCenter after failover..
• They can manage Hosts at remote from site B by directly connecting to Hosts at remote sites...
• Setup 3rd stand by vCenter site. Restore Site A vCenter configuration on standby Site. This would allow customer to manage 100+ sites via vCenter.

Does anyone know of a way to do this?

[sjesse]

The option that comes to mind at the moment is have 2 vcenter at production site a and once vcenter at production site b. 1 vcenter is just for SRM and managing the management vms you want to fail over. The second one, which you can protect with SRM, that would fail over to site b. If you can't stretch layer 2 between the datacenters, when if fails over someone would have to login to the vcenter appliance and update the ip, and make sure dns is updated for the new ip.

Edit the DNS and IP Address Settings of the vCenter Server Appliance

I'm not sure if SRM scripts would change this or not, technically these are Linux and vmware tools is installed, so it may work. I'd run this by support through and make sure this is supported, or maybe someone else would come up with a better option.

[pwilcox]

Okay so we would SRM's "Shared Recovery Site" topology.

Protected Site A vCenter 1 appliance would support the management VM's. Management VM's protected by SRM utilizing vSphere replication to Recovery site B vCenter 3.

Protected Site A vCenter 2 appliance would manage the 100+ Hosts and clusters at the remote sites. vCenter appliance protected by SRM utilizing vSphere replication to Recovery site B vCenter 3.

In the event of Disaster at Site A.

• Management VM's would be failed over using SRM to site B. Assume DNS configured correctly so that management VM's are accessible.
• vCenter appliance protected by SRM would be failed over to site B. IP address would need to be updated on appliance prior to bringing online. Assume DNS configured correctly so that remote sites would be able to communicate with vCenter 2.

Failback

• Management VM's would be failed back using SRM to site A.
• vCenter 2 appliance would be failed back using SRM to site A.

Is this a supported solution?

[sjesse]

I think so, but I'd get confirmation from GSS. There was a white paper release for Vmware View 5, and before Cloud pod architecture was released, it looks like moving the vcenter server in this method was the recommended method

https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/techpaper/vmware-view-vcenter-site...

its where I remembers the idea from.