Why does host file system and management softwares show different space usage than ONTAP?
Applies to
- Data ONTAP 7
- FlexPod
- SAN
Answer
When turning off LUN space reservations, data inside of LUN is not represented by space df shows as taking up.
- Following example uses a freshly created volume that contains a single LUN for illustration purposes. The volume is space guaranteed on the aggregate and has factional reserve setting at 0%:
storage> lun show -v
/vol/space/space 86.3m (90478080) (r/w, online, mapped)
Serial#: nBLskoOLyq/2
Share: none
Space Reservation: enabled
Multiprotocol Type: windows
Maps: space=0
storage> df -h
Filesystem total used avail capacity Mounted on
/vol/space/ 174MB 86MB 87MB 50% /vol/space/
/vol/space/.snapshot 0TB 0TB 0TB ---% /vol/space/.snapshot
df -h command shows the amount of space in the volume that LUN is using plus reserved space for LUN's total size.
- Disable space reservation:
storage> lun set reservation /vol/space/space disable
storage> df -h
Filesystem total used avail capacity Mounted on
/vol/space/ 174MB 2536KB 171MB 1% /vol/space/
/vol/space/.snapshot 0TB 0TB 0TB ---% /vol/space/.snapshot
With space reservation disabled on LUN, df displays the amount of used space within LUN on that volume.
- When space reservation is turned back on and a 30.6MB (30668KB) file is written to this LUN:
storage> lun set reservation /vol/space/space enable
storage> df -h
Filesystem total used avail capacity Mounted on
/vol/space/ 174MB 86MB 87MB 50% /vol/space/
/vol/space/.snapshot 0TB 0TB 0TB ---% /vol/space/.snapshot
The storage system shows the same as when there was no data inside of the LUN.
- If reservation is turned off again:
storage> df
Filesystem kbytes used avail capacity Mounted on
/vol/space/ 178176 33276 144900 19% /vol/space/
/vol/space/.snapshot 0 0 0 ---% /vol/space/.snapshot
~30688KB file is now added to the previous 2536KB.
- Turn reserve back on and delete the file:
storage> df -h
Filesystem total used avail capacity Mounted on
/vol/space/ 174MB 86MB 87MB 50% /vol/space/
/vol/space/.snapshot 0TB 0TB 0TB ---% /vol/space/.snapshot
- Once the file is deleted, the host file system shows the space is available; however, Data ONTAP shows the same 32MB is being used. This behavior occurs because the space used by the file is still allocated on the LUN. The host file system is able to overwrite this space as needed.
NetApp OnCommand System Manager reports different space usage of a LUN
- Run the following command to ensure OnCommand System Manager is reporting the same space used as Data ONTAP CLI:
::> lun show -vserver VSERVER_NAME -path COMPLETE_LUN_PATH -instance
Different VMFS-LUN size reported by VMware Web client than ONTAP
- VMware Vsphere 5.x does not automatically release the blocks after any VMs are deleted or moved. Hence, the blocks are not marked as free to the storage controller; although from the VMware client itself, it recognizes the blocks as free.
For RDM space discrepency, the following additional steps can be tried if the above action plan does not help:
- Increase the size of the RDM LUN by 50 GB perform rescan on the ESXi host level and rescan within the guest OS.
The second task will have to be performed if the difference of the free space is not resolved by a first step.
- Power down VM, remove RDM disk from the VM and re-add this RDM back to the VM under the same SCSi controller position. (example - 1:5)
- Power up VM and check if the free space on the disk is matching on the guest OS and the storage array side.
Perform the above task for all RDM LUN's that are reporting different free space on the guest OS and storage array.
Additional Information
N/A