I have created a NFS datastore on my Filer and presented this to my ESX server (3.5) and migrated some servers into it. However the datastore reports back that it still has the same free space as it did before the migration?? The data usage on the Filer reports back correct but not in the VI client (2.5).
This is the first time i have used NFS datastores with VMware as we currently use VMFS LUNS over iscsi. I haver tried refreashing and logged in and out of the VI client but still no joy.
Isn't the volume presented as NFS share thinly provisioned by any chance (space guarantee set to none)?
If that's the case and its size is bigger than the containing aggregate free space (say 600GB volume within aggregate having 500GB available), then on the host side you will see 600GB datastore with 500GB free space.
So, say you've moved a vmdk file from a LUN within the same aggregate to the NFS share & deleted the LUN - the amount of available space will remain unchanged.
This is strange, the volume isn't thinly provisioned nor is it set for De-dupe. The Filer reports back correctly stating that 47.4GB of data is used in the volume but the VI client reports back no change!!!
Yep as the last post mentions, you will have to refresh the storage view in the VI client. Its not that dynamic if at all and has always needed a bit of a kick... I guess this should get better the more the vStorage api is used with VC in the future and the integration is tighter between VC and the storage layer.
Personally I would only use the VI client storage view (with respect to NFS) ever as a rough guide, in fact I never really look at it unless I'm mounting a datastore always using the filer tools or the service console instead.
Yes its not great that its not good at updating but I would personally prefer ESX to use its CPU cycles for real work!