I have a pair of VM's that are running an MS SQL cluster (using MSCS failover clustering). The VM's are set up as outlined in http://pubs.vmware.com/vsphere-50/topic/com.vmware.ICbase/PDF/vsphere-esxi-vcenter-server-50-mscs-guide.pdf and the clustering is working well. The problem we have is that whenever the active node VM is reconfigured, such as adding a NIC or a USB device, the MSCS cluster reports an event that the raw device mapped-LUNs have failed and the cluster initiates a failover -- even if no changes were made to the RDM hardware or anything related to storage. If i reconfigure a non-cluster machine I do notice that the disks flicker briefly as if there's some type of momentary rescan, but for whatever reason, the MS cluster service sees this is the disk being removed from the machine which causes SQL server to immediately stop and failover to the other cluster. I've confirmed that changing hardware parameters (like a NIC's portgroup) does not cause this to happen, but any change in the hardware line up (adding a NIC, removing a NIC, modifying USB devices, adding an unrelated disk), will cause the issue to occur. The most specific event Windows reports is "The disk has failed" in reference to the shared storage. I can reproduce the problem on either node as long as they're the active node. Any help would be appreciated. The hosts are running ESXi 5.0 Update 1 and are both up-to-date.
Simon