Cluster disk resources fail or go offline after paths to available virtual disks are disabled

Products

• VMware vSphere ESXi

Issue/Introduction

After you disable the paths to all available virtual disks, including boot disks and cluster disks, in an active Windows Server 2008/2012 cluster node, the cluster resources fail or go offline.

Environment

• VMware vSphere ESXi 6.0
• VMware vSphere ESXi 5.1
• VMware vSphere ESXi 5.5
• VMware vSphere ESXi 5.0

Resolution

To trigger a disk resource failover from the active node to a peer:

 

In vSphere 5.0:

1. Modify the .vmx file for the Windows Server 2008 SP2 and later cluster nodes and add this line:
   
   scsiX.returnNoConnectDuringAPD = "FALSE",
   
   Where X is the shared SCSI bus controller ID number.
    
2. On the active cluster node, set the Maximum restarts in the specified period property for all IP address resources to 20, so that the failover happens under APD conditions.

In vSphere 5.1/5.5/6.0:

1. Modify the .vmx file for the Windows Server 2008 SP2 and later cluster nodes and add these lines:
   
   scsiY.returnNoConnectDuringAPD = "TRUE"
   scsiY.returnBusyOnNoConnectStatus = "FALSE"
    
2. If the resources do not fail over after making the above changes, then modify the .vmx file as below:
   
   scsiX.returnNoConnectDuringAPD = "FALSE"
   scsiY.returnNoConnectDuringAPD = "TRUE"
   scsiY.returnBusyOnNoConnectStatus = "FALSE"
   
   Where:
   • X is the shared SCSI bus controller ID number.
   • Y is the boot device SCSI bus controller ID number. By default, Y is set to 0.
3. These settings are also added to the .vmx file in ESXi 5.1/5.5. This prevents the ESXi host to stop the virtual machine when an All Paths Down (APD) situation is encountered.
   
   msg.autoAnswer = "TRUE"
   answer.msg.hbacommon.askonpermanentdeviceloss = "retry"