Dear Community,
I am managing to remove NFS datastore from ESXi 6.0 servers and task lasts for 38 minutes to finish :smileyshocked:
Did you guys see this issue on your infra?
Regards,
Fred
Hi!
It's not a normal behaviour.
Did you remove it using vcenter web client? Or just on standalone host?
How many hosts were connected to the datastore?
Hi Finkiez,
I removed NFS datastore from Orchestrator Workflow, doing it 1 ESXi server by 1
NFS datastore is mounted on many many hosts at the same time. I guess this could be the root cause....
This can be issue that you remove from ESXi hosts one by one.
If you try remove NFS datastore using web client - does it work fast?
You are seeing slow unmounts (which precede "datastore delete" actions). Slow unmounts are almost always caused by Network issues (NFS Server unreachable) or by hosts locking the datastore in some way. Can be HA using a NFS datastore.
A good practice is to disable HA in advance, or de-select NFS datastores used as HA heartbeat datastores. I'd rather briefly disable HA so avoid problems as sometimes, despite NFS not being selected for heartbeating anymore, they are still actively used by HA.
After all the mounts are gone, and this can be done in one go via the wizard, enable HA again.
It helps looking at the vmkernel.log file in realtime to see what ESXi is doing (and not succeeding at). Select an ESXi host do to an unmount and run this command in an SSH session while you unmount:
"tail -f /var/log/vmkernel.log | grep -i nfs"
NFS datastore is not configured for datastore heartbeat
I will unconfigure HA to see if unmount is faster
I want to notice you that unmount operation is more and more faster if I do it again and again....
Thanks for helping guys!
