You will need to unmount the NFS datastore and then re add it.
"unmount" sounds like a Service Console concept. Is that really what you mean? Login to the SC and unmount the volume?
Or do you mean "Delete" and then re-add from the VIC?
Once I do this, and assuming I use the same name, will all snapshots/VMs using the datastore simply connect using the name when I bring them up?
Actually - maybe I have a different problem.
I notice that *all* my NFS datstores are "inactive", even those whose IP addresses have *not* changed!
I checked and I did change my VMKernel IP address and router properly (using VIC/vCenter in Configuration/Network). What else might be wrong?
Do I need to issue a command on the Service Console itself?
Have you updated your DNS server? Make sure that you can resolve the NFS server names from each ESX(i) host. Ping etc.
You can try restarting the management agents from the console.
I figured this out, after researching a few other threads.
For posterity, here was the issue:
Once any datastore's IP address is unreachable, the ESX cannot mount any of them, as reported by "esxcfg-nas -l".
Also, as long as any VM (even one which is not running) has a reference to a datastore, you cannot delete it while ESX is running.
So, my solution was:
- Shut down all VMs (via the VIC)
- Put the server into maintenance mode:
# vimsh -n -e /hostsvc/maintenance_mode_enter
- For each datastore whose address has changed, delete the old, and re-add it with new address (and exact same label):
# esxcfg-nas -d <label>
# esxcfg-nas -a --host <new_name> --share <remote_nfs_share_path> <label>
- Mount the datastores
# esxcfg-nas -r
(this will fail if any datastore hosts are still unreachable, or will complete silently if all goes well)
- Bring the server out of maintenance mode:
# vimsh -n -e /hostsvc/maintenance_mode_exit
I hope somebody else finds this helpful (as I would have :-).