NFS clients experience mount failures on specific ESXi hosts within a vSAN cluster. Diagnostic symptoms include:
showmount -e localhost from within the vsfs container on the affected File Service Virtual Machine (FSVM) returns: RPC: Timed out.ganesha.nfsd process appears in an Sl+ (Interruptible Sleep) state when checked via ps or top inside the container.vSAN 8.x
vSAN File Services
The Ganesha NFS daemon (ganesha.nfsd) enters a hung state while processing RPC requests. This is typically characterized by a broken pipe in the transport layer, preventing the daemon from responding to new mount requests or I/O operations from clients.
In the ganesha.log (located inside the FSVM container at /var/log/ganesha.log), the following error repeats: rpc :TIRPC :svc_ioq_flushv() writev failed (32) (Broken Pipe)
This indicates that the daemon's internal I/O queue for RPC services has encountered a terminal failure on a socket, causing the service thread to hang vSAN FSVM Mount Failure – “No such file or directory” Error Observed When Mounting File Share
To resolve the hung state and restore NFS services on the affected host, the vSAN File Service (vsfs) container must be restarted.
vsfs container ID:vsfs container to reinitialize the Ganesha daemon:showmount -e localhost returns the list of exports successfully.