Consumed Space Not Changing LF OS VSA/P4000

[dallasbr]

 Hello,

On our P4000 San and our VSA San the CMC shows that we are using more space than we actually are. For example if I go in and delete a 10GB file off of either SAN, the SAN never shows less consumed space.

This is coming up because I am getting a warning that I have exceeded the 95% threshold on the VSA Volume, so I deleted a pile of data, and the consumed space never drops. It still shows the exact same amount of consumed space. Anyone know why? Or how to get this to accurately update.

Thanks

dallasbr

P.S. This thread has been moved from Storage Area Networks (SAN) (Small and Medium Business) to HP StoreVirtual Storage / LeftHand. - Hp Forum Moderator

[CalvinZito]

I assume you're using thin provisioned volumes.  When you delete a file, that space isn't returned to a "free space" pool.  There has been work over the last couple of years by VMware and partners (like HP) to use the SCSI UNMAP command so that an array can tell the OS that the space is now available and it shows up as free.

Currently StoreVirtual doesn't support UNMAP.  If you are using VMware Thin Provisioning, you can shrink VM disks to recover space but not reclaim space from deleted VMs. 

Maybe someone else has more insight into how from the OS side you can return the deleted file space to the free pool.

[dallasbr]

Thanks for the reply. The storage is FULL provisioning. The VSA is connected to a Server 2003 physical machine with one Volume using iSCSI. Very Simple setup. Just seemed odd to me that the SAN would not update the consumed space, because essentially I have no way to get rid of the error message.

[oikjn]

there is currently no space reclaimation function on the SAN so once space is allocated the only way to free it up is to delete the LUN (or snapshot) using that space.

I don't know why you are using full provisioning and should probably switch that over to thin, but the SAN is behaving exactly how it is supposed to.  you are running low on ALLOCATED space, but you have allocated it all w/ your large THICK volume...  I'll give you it would be nice to turn off the warning if you wanted, but your space constraint "problem" is completely because you went with one giant thick volume.

I woudl suggest you NOT run one giant thick volume and brake it down into a few 1TB or 2TB LUNs.