How do you provide site (data centre) redundancy for vcenter?

[abugeja]

Hi,

I need to provide full site (data centre) redundancy for our vcenter environment. Currently we have ESXi hosts located at both our data centres they then hook into a single instance of vcenter.

If the data centre goes down where vcenter instance is located we are unable to see how the environment looking. Unsure if its all VMs impacted or only a selected few. We could login to each of the ESXi clients but we are a large environment so its not best and takes time.

from vcenter 5.5 u3 you can use windows clustering for vcenter but this is only supported on virtual machines and it needs to be at the same site as it uses RDMs

Surely other people have been asked to provide a true HA for vcenter?

[virtualg_uk]

Hello,

If you just need HA for your VC then all your supported options are here:

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1024051

I hope this helps.

[abugeja]

No as it doesn't resolve site redundancy for vcenter. Yes you can have a vcenter at both sites but what happens if vcenter goes down at one of the sites. You are then unable to manage the environment.

Why cant you failover the VMware services much like a SQL GEO Cluster to the other site and you are backup and running and then you have visibility of the site that is partially down.

[ThompsG]

Hi abugeja,

I would suggest re-reading the link sent by grba as it describes your options in regards to HA for vCenter. While it is currently not supported, if you use SQL AlwaysOn for the database and Windows Failover Clustering for vCenter then you could possibility achieve site redundancy for the solution.

Another way to do it would be to have something like a replicated management cluster which you could failover to your DR site if required. This could be done via vSphere replication or array based options. Documented process and you would be away again.

Of course if you are wanting something completely automated and detects site failures then you are correct - there aren't many options. Probably why VMware removed vCenter Heartbeat - gathering there wasn't the need for such a product especially with the push to the vCenter appliance.

Kind regards.

[virtualg_uk]

Correct re replication etc, and also HA can protect the VC also FT capable in v6.

Key really is to backup and if there are issues that cannot be fixed via a VMware support request, then you have the backup available to you.

Yes, the VC is required for certain management functions but most important ones can be completed directly by logging into the vSphere client to each host. Not ideal but this method is available until the VC issues are resolved.

A combination of backup and replication is the best option here as then you can recover in case of site failure and in case of VC corruption / OS issues etc