Local datastore not accessible after replacing local storage controller on ESXi host

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms: 

• After replacing a local storage controller, the local VMFS datastore now appears to be inaccessible in the "vCenter UI > Hosts > Datastores or vCenter UI > Hosts > VM > Datastores".
  
  
  
  
• VM's located on the datastore may also appear as inaccessible. 
• The physical storage device appears to be attached, but not consumed in the "vCenter UI > Host > Configure > Storage Devices":
  
  
  
  
• From an ssh session to the host, the datastore does not appear in the /vmfs/volumes directory. 
• In some cases it is observed that the local datastores are not visible to the ESXi host. This can be validated using the commands "esxcli storage filesystem list" and "esxcli storage vmfs extent list".

 

 

Environment

VMware ESXi (All versions)

Cause

• The datastore now appears to be a LUN replica/snapshot copy of the original disk.
• This is due to the controller replacement.
• When you replace a controller, the disk may report a new device ID ( t10.*, naa.*, or eui.*) or LUN ID (reflected in the vml.* ID) or both, triggering the snapshot LUN detection. 

Verification:

• To verify if the LUN is being detected as a snapshot, see the following KB: Troubleshooting LUNs detected as snapshot LUNs in vSphere

Resolution

• The datastore will need to be resignatured. See KB: Troubleshooting LUNs detected as snapshot LUNs in vSphere
• After resignaturing, you may rename the datastore back to its original name. See KB: Renaming a datastore
• Any inaccessible VM's that were located on the datastore, will need to be unregistered, then re-registered.  See KB: Add or Register a Virtual Machine (VM) in vCenter Server