Solved: Re: LUN SPACE CONFUSION

NAYABRASOOL · ‎2012-09-07

Hi All,

I have got an alert for my volume threshold reached as the space left on volume was only 12MB out of 107MB i have lun on top of this Volume which is of size 102MB, i have conveyed the client about this but he was telling the lun what he is using on the windows side it was showing 61.1 MB free , so could anyone please explain about this.

Thanks,

Nayab

GARDINEC_EBRD · ‎2012-09-07

I'm sure there are KB articles about this, but don't have time to go hunting right now, sorry.

As a quick example, if you delete a load of data from the client-side (eg NTFS) the client marks the blocks as free, as opposed to physically zero'ing out the data, right? Down at the storage level WAFL has no way to know these blocks have been deleted, so when you write more data to the LUN it will consume new blocks in the volume. Typically, as a LUN ages, you will find the NetApp side will show the LUN at, or close to 100% full, but the clients filesystem may still have plenty of space. This is by design, and often not a problem - although it looks a bit odd at first. Check out Snapdrive's Space Reclaimer feature if using Windows - this will reclaim those free blocks at the WAFL end if required. Docs around this will also explain why you see the difference in more detail than I've done so here.

Regards,

Craig

View solution in original post

pinrell · ‎2012-09-07

Hello

Could be several things,

Mainly snap residing within the volume:

It is a Lun thinprovisioning??,

Has fractional_reserve =???

Reserve for Volume Snap?

GARDINEC_EBRD · ‎2012-09-07

I'm sure there are KB articles about this, but don't have time to go hunting right now, sorry.

As a quick example, if you delete a load of data from the client-side (eg NTFS) the client marks the blocks as free, as opposed to physically zero'ing out the data, right? Down at the storage level WAFL has no way to know these blocks have been deleted, so when you write more data to the LUN it will consume new blocks in the volume. Typically, as a LUN ages, you will find the NetApp side will show the LUN at, or close to 100% full, but the clients filesystem may still have plenty of space. This is by design, and often not a problem - although it looks a bit odd at first. Check out Snapdrive's Space Reclaimer feature if using Windows - this will reclaim those free blocks at the WAFL end if required. Docs around this will also explain why you see the difference in more detail than I've done so here.

Regards,

Craig

NAYABRASOOL · ‎2012-09-07

Thumps Up !!!!! Craig Thanks a lot ,

As i am new to Netapp i don't have idea about this stuff , can you provide me the link to that KB article about which you are speaking please.

Regards,

Nayab

saranraj456 · ‎2012-09-07

Provide the o/p of lun show -v

SNETAPPUSER · ‎2012-09-07

GARDINEC_EBRD is correct. Use SnapDrive for windows to reclaim free space from thin provisioned LUNs from a Windows box. If you encounter the same problem with VMWare, you have to do it manually using SCSI UNMAP commands in ESXi shell. Here is a KB article for that. http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2014849 (Note that, your storage system should support VAAI capability.)