UPDATE 02.15.08: I updated the last line of the VMX file in step 4 (Thanks for the heads up Jim!)
I have a need to build an Exchange 2007 single copy cluster on Server 2003 in VM environment. This is something I've done before with Microsoft Virtual Server, but haven't had the opportunity to do it with VMware Workstation (OK, I had the opportunity, but never needed to until now). Since VMware Workstation is a paid product, I assumed clustering and setting up a shared SCSI bus would be included and pretty straightforward. Wrong. No options in GUI exist for this, and the VMware help files and knowledgebase provide no assistance. I love VMware, but come on, no easy way to set up a cluster in a product that is meant for lab testing various configurations and applications?
Anyway, after poking around on the web for a while, I was able to compile a working method for setting this up. Here are the steps I took:
-
Create two VMs with a single disk on SCSI bus 0, install Windows, get updates, etc.
-
On one VM, add two new SCSI disks (one for data, one for quorum) on SCSI bus 1 (I put them on SCSI1:1 and SCSI1:2), make them independent, persistent, and fully allocate the drive space so that the VMs don't fight over expanding/managing the disk file size.
-
On the second VM, add the two already created SCSI disks on SCSI bus 1 (again, I put them on SCSI1:1 and SCSI1:2), and again make them independent and persistent.
-
Edit the .vmx files for each VM and add the following lines:
scsi1.sharedbus = "Virtual"
disk.locking = "false"
diskLib.dataCacheMaxSize = "0"
diskLib.dataCacheMaxReadAheadSize = "0"
diskLib.dataCacheMinReadAheadSize = "0"
diskLib.dataCachePageSize = "4096"
diskLib.maxUnsyncedWrites = "0"
-
Fire up the VMs and both should have two disks ready for formatting and clustering.
Admittedly, the lines added to the .vmx file came from Geert Baeke’s Blog, specifically his posting on clustering in VMWare Workstation 4.5 and higher. Also, he has another good post around clustering in VMware Workstation using iSCSI. Trying that out is already on my “to do” list.
As a side note, when you complete botch the configuration of your Windows Cluster (like I did) and want to start over, use this command from the C:\windows\cluster directory on each node, then reboot the node:
cluster node <node_name> /forcecleanup