HPE Storage Users Group

hi,
We have a 3PAR F200 (InServ 3.1.2 MU1) and vSphere 5.5.
I executed a reclaim on each vmfs (esxcli storage vmfs unmap).
The task in the InServ continues in the background: showeventlog -debug -oneline -msg "reclaim"

Once completed, since the GUI and showvv commands (showvv -r vvName & showvv -cpg cpgName -r -p -type base) I see the recovered space.
The problem is that since the GUI or command showspace (showspace -t -p r5 -devtype FC -ssz 4) RawFree and UsableFree values are identical to before.

I forgot the recovery operation by the CPG.

Thank.

_________________
2 F200, waiting 2 7200c

Compactcpg?

Yes, a task is scheduled by default (Sunday) by the InServ.
When I submit "compactcpg -trimonly -f cpgName"...
Compacting cpgName
CPG cpgName already compact

No recovered space.

_________________
2 F200, waiting 2 7200c

The space will first be returned to the VV so will still have multiple 16K pages allocated in the underlying LD. This space will only be returned to the underlying CPG if 128MB's of contiguous pages exist, which may require multiple defrags to run. At which point it will then be returned to CPG and can then be returned to the system as a whole via a compact CPG.

Thank for this explaination.

So I have to wait ? Can I run a defrag to have spaces of 128 Mo ?

_________________
2 F200, waiting 2 7200c

The only way to force an immediate defrag today (do a search on the forum for other discussions on this "space reclaim") is to tune the VV from a TPVV to a FPVV and then back to a TPVV for which you need 'DO'. Obviously you need space to do this, so whether it's practical will depend on available space, how much you can potentially reclaim and how quickly you need the space.

Not to mention how busy your array is since DO is a low priority process (not background like scrub and some other functions, but the array does it's best to make sure it's minimally impacting), if your array is very busy DO can take a very, very long time. For us moving ~100TB took over 7 weeks.

Thank you for your answers.
I read pptx posted by JohnMH

I do not have enough space for the conversion of the VV (vmfs). I could create a new VV and delete the old one.
We will migrate to 7200c and I wanted to understand better manage.

_________________
2 F200, waiting 2 7200c