Collectives on same physical virtual server hosts

Discussion created by RJKSolutions on May 7, 2009
Latest reply on Jul 8, 2009 by andreas

I came across a statement in the HA & PI Server Replication document:


Note: The PI Server can be run in Virtual Machine (VM) environments provided by
Microsoft Virtual Server or VMware. Incorrect use of virtualization may result in a
single point of failure. If you decide to run the PI Server in a VM, we highly
recommend running the servers in a PI Collective on different VM host hardware.


Curious, is this just down to point of failures (i.e. if the host goes down then the collective members on the VMs go down) or is there a hidden technical reason?




Came across this OSI article on virtualisation, apart from data storage/loads anything else to keep an eye out for - in terms of PI?  Along the lines of system resources, virtualisation host considerations..?