I've created a 500GB volume and migrated 3 servers into that volume. vCenter sees that volume with 509GB provisioned and 22GB free space left.
Nimble reports that the volume is using 237GB, so I should have over 250GB of free space on that volume. However, since vCenter doesn't see what Nimble sees I cant move any more servers onto that volume. vSphere reports "Insufficient disk space on datastore xxx".
VMWare ESX will write to it's datastore as if it's a regular disk... the Nimble array will compress that data in-line (with no degradation to performance) so although ESX is reporting it's stored 500GB, the array has compressed that into the 237GB (saving you the space in return)
If you've not done so already register the Nimble vCenter Management plugin as this will allow your vCenter server to view the Nimble volumes (plus much more)
Feel free to post back if you need more info or have more questions
This had me baffled for a moment until I realised exactly what the KB article outlined. Probably a reason for some SAN admins preferring not to use thin provisioning. The simplicity of reporting available usage and space on the Nimble GUI makes me for one comfortable with thin provisioning and compression. The other thing that threw me was the SCSI UNMAP. Have a quick read about this and you will have a little more clarity about what one system says versus your Nimble array.
