Network and Storage Protocols

The real "free space"


I didn't wanna bring up an old thread

But there is some good info in there..  Curious tho, how would I caculate the "real free space" in an aggr.  If we have a mix of volume guarantee volume and none, with only NFS and CIFS data on an aggr,  Could we not run the aggr up higher then 90% should we have a real free space in some of these volumes largely unused?

If we have 1T amount 10 vols at guarantee volume and only 10G in each, that's 100G, or 10% used. so that leaves with 900G of space "Free" right. running the aggr up to 90% means we may only really useing 80% of the aggr in this case.



I think you can use df -g -A to view aggregate space usage, including free space; and df -g to view volume space. Thanks,   -Wei


That's not what I'm looking for.

Real aggr free space. After you take into account the free space within volumes with guarantee = volume


Ahh...well if your Volume is Space reserved at Volume level, then are you looking at free space at Volume level or Aggregate level ?

If you have volume reservation enabled then as said above 'df -A' will give you space reservation at Aggregate level.

Consider : if Aggregate has 100GB. and you created a Volume with 50GB. ( Space reservation enabled). You have free space at aggregate level 50GB.

This is because you have reservation enabled, hence space is carved out the moment it it created. Irrespective of the free space in the Volume.

This is exactly the opposite if reservation is at = none level.

Hope that helps !!!


try aggr show_space, lots of useful information there.


I guess it depends on your definition of "real" free space.Surely a LUN with 50% free space within the OS would also constitute free space? If you thin provision all your volumes to "none" you would free up a lot more space than if you just set them at "file".

I don't think there is a simple command to show you all the info, but you could probably do something with a quick script or Excel. Chuck all the volume sizes into Excel and you could easily show the free space in volumes + aggregate.

If you set all the volumes to guarantee=file, the aggregate free space would then give you your "real" free space. If you scripted this, it wouldn't need to affect the system at all.

If you set the volumes to guarantee=none, the aggregate free space would then also show you unallocated space within your LUNs also. If you are using the SnapDrive space reclaimer, this would be more realistic than if you aren't (as deleted blocks from Windows may still be taking space on the NetApp).