I am reading the vi3 vm and mscs pdf. It says that the boot disk for the VM's need to be on local storate. I only see SAN storage when trying to create the VM. I did not create any local VMFS partition on the ESX servers during install.
The extra space was used as a /vmimages partition for storing ISO's. Am I f'ed?
Let me first preface this by saying, I am no expert in MSCS in ESX3. I have configured a test cluster across physical boxes using SAN storage for the VM's boot disks. Most of the MSCS feature set worked in regard to being able to initiate cluster resource moves manually. Simulating failures produced some interesting results. Simulating network failure, it looked like the cluster failed over correctly. Simulating loss of power/windows crash of the active node did not generate an automatic failover response.
I had tested that configuration before reading that same document. I did have local storage space and was debating setting up the cluster in the prescribed manner. I have instead decided to look into DFS (replicated file shares) to provide the service that I have been asked to configure. There were quite a few concessions that I would have needed to make in my current ESX environment based on the mscs pdf that I just did not want to make.
As far as partitioning goes, the only way that I know to change that would be to reinstall ESX on that host and make sure that a local vmfs partition is made. If you have Vmotion and multiple physical hosts, that would not be a terrible experience.
Hope that helps,
Thanks Ken. I figured I would have to reinstall ESX. We do have enough blades here to take one esx host down at a time. The thing is is I am the vmware consultant they brought in to do this correctly. I didn't realize they were going to be using mscs so I didn't set this vmfs partition up.
I just don't want to look like an ass for "forgetting" to create it. Although all my docs were signed off by my manager and director...
I'm aware that vmware supports MSCS only when the OS vdisk is on local storage. However we have been running a PRD DB cluster on SAN storage with shared RDM LUNs since ESX 2.5.x and now on 3.0.1 without problems.