VCS cannot take virtual machines offline on NFS datastores when the NFS server or NFS shares are inaccessible. An NFS datastore can become inaccessible due to any of the following reasons:

-     The NFS server crashes, which leads to the NFS share getting unexported from the NFS server.

-     The ESX hypervisor loses network connectivity with the NFS server.

In these situations, VCS can detect the fault for the corresponding VMFSVolume resource.

 
Depending on policy settings for handling faults and failover, VCS tries to take other resources in the service group offline. When VCS tries to take the ESXVirtualMachine resource offline, the resource can get stuck trying take the virtual machine offline, even with the force option. VCS eventually marks the resource state as:

UNABLE_TO_OFFLINE | Admin Intervention


This behavior is noticed only with NFS datastores. It occurs because the ESX Server marks the virtual machines as inaccessible and rejects all (graceful and force) operations on virtual machines that have their boot image on the datastore. This is a VMware limitation. Symantec is working with VMware on a possible resolution or workaround for this issue.