Vcenter Heartbeat End of Availability: Thoughts and Opinions

It appears that VMware is going to do away with this product for high availability and I can’t say I am surprised.  You can read about why they are doing it and the release at this link:

We had attempted to use this HA implementation even though our VCenter was already a virtual machine using VMware’s HA.  We had been under the impression that VCenter Heartbeat would provide no downtime failovers so we could do maintenance on one windows server while maintaining 100% availability for our VCenter.  Unfortunately, this was not the case with this product.  It used NeverFail’s HA technology which in essence was a stopping of the original service and  starting of the new service/failover of VIP on the secondary server.  This of course could take anywhere from 1-10 minutes to actually have the VCenter service available depending on your inventory size.

So, we ended up scrapping the software and using VMware’s builtin HA.  This provides the same level of availability.   If you currently have a physical VCenter, VMware now recommends to virtualize your VCenter. I also recommend this methodology.  That being said there are 2 possible routes I would go to do this.

  1.  I would create a management cluster so you know which host your Vcenter will be on in case it goes down.
  2. If you can’t create a management cluster, setup a DRS rule group to allow  your Vcenter to only run on 2 hosts so that you know how to get to it in case of failure.

All this being said, what I would love to see from VMware is a true fault tolerant VCenter where you can have 99.99% uptime where you can fail over services to a secondary server without them going down so you can do any maintenance on the OS.  VMware any future solutions for this?