VMware just let it slip that they plan to announce a new HA solution specifically for their vCenter management server at VMworld Europe next week. Right now the vCenter management system is a single point of failure in a VMware virtualization deployment. If the vCenter server goes down your setup then runs in a degraded state until you can get it back up and running. This new solution is really an active/passive clustering setup between two vCenter instances. Consultants have been doing this type of setup in field deployments using third party clustering software for a while now. So this is really just a bandaid to try and reduce their single point of infrastructure failure pain point. Now customers will have to setup and maintain not one but two backend vCenter instances and also now add in the complexity of a clustering solution on top of that. Word is that it’s a third party solution, so only time will tell how integrated this all is.
The end result of all this is a higher cost of deployment and more complexity in a customer’s VMware deployment. Go ahead and add an extra few days to your VMware deployment schedule for this. And upgrading this whole setup will be loads of fun every time a patch gets released.
Plus not only do you have to deploy a second instance of Windows Server to run this on you also have to pay VMware about $10k for the privilege of fixing a flaw with their core management infrastructure.
Citrix XenServer on the other hand includes an elegant fault tolerant peer-to-peer management layer at the core of its management architecture. Not only do we not have a single point of failure in our core XenServer management infrastructure we don’t charge extra for it and don’t require a dedicated management server. Plus it’s so easy to use you don’t even have to do anything to take advantage of it, it’s all automatic, and under the covers. No databases to cluster, no extra machines to purchase, no extra windows servers to deploy, just simple and easy. 10 minutes to Xen.