Several times I have come across the situation when the NFS datastore on the VMWare ESXi host becomes unavailable / inactive and greyed out in the host’s storage list. However, the NFS storage stays available on the network level. Usually, it can be solved by removing the NFS storage and then adding it back.
However, it happens that when trying to remove such inactive NFS storage from the ESXi host configuration, it returns an error:
Based upon the error, the NFS storage cannot be removed due to the Storage I/O control (SIOC) is enabled. If you disable it, another error appears:
Solution
To properly remove the NFS storage from the VMWare ESXi host, you must directly connect the ESXi host server via SSH (you can enable SSH from vSphere Client).
Display the list of the NFS storages in the system:
esxcli storage nfs list |
A similar command in ESXi 4.x looks like this:
esxcfg-nas -l
|
To unmount the NFS storage:
esxcli storage nfs remove –v datastore_nfs02 |
In ESXi 4.x command is as follows:
esxcfg-nas -d datastore_nfs02
|
If the NFS datastore isn’t removed from the vSphere Client, click the Refresh button in the ESXi storage section (Configuration -> Storage).
In my case the NFS storage was bisy with one of VM. Via thick vSphere Client I so it in Inventory “Datastores and Datastore Clusters” at “Virtual Machines” tab. Although there was no devices placed there at that moment and in fact of that the VM’s properties didnt content storage in storage list. But earlier there was mounted a CDROM to iso file at (i used this storage for delploy ), and there was a snapshot (i dont use snapshots in my work, but it was autogenerated with clone operation and in unknown reason didnt removed automaticaly). Then I deleted snapshot and occured VM covered from earlier tab, and then I abled to unmount storage.
Clearly, in except of using so hard manipulations (and in commonly, not preferable), in first step we should release storage from any obvious relations.