df on filer shows larger amount of used space for volume than Windows host

paul_w_jackson · ‎2011-03-09

I have a flex volume with the following options:

Volume State           Status            Options
BOSEQUCO01_Vol_2 online          raid_dp, flex     create_ucode=on,
                                redirect          convert_ucode=on,
                                active_redirect   fractional_reserve=0,
                                                  snapshot_clone_dependency=on

It has 1200GB total, 453GB used according to a "df -g" run on the filer. This volume is exported as a LUN to a standalone Windows 2008 server and when I get properties on the drive, I see that it's only using 381GB. What's going on with the 72GB that the server doesn't see? Is this common behavior for LUNs? I had this same behavior the other day and ran a SnapDrive Space Reclaimer job and it was able to eliminate the "ghost" space. However, now when I run the reclaimer it says that there is no space to reclaim.

Thanks,

Paul

columbus_admin · ‎2011-07-07

Most likely if this isn't a whitespace issue...it is a snapshot issue. What does a df -g show is the reserve over, and what are your LUN settings? Is the LUN guaranteed, thin provisioned? lun show -v should help get most of that.

- Scott