HPE Storage Users Group

All,

I've been trying to get a 3par 7400 array to release vacated space all day long with no luck.
I ran vmkfstools -y XX (numbers from 65 to 99) and the console displaced creating balloon file etc, yet no space was reclaimed. So far 8 hours has passed with no change in space. I also tried creating a zeroed out flat file with vmkfstools -c <size g> -d eagerzeroedthick zerofile, and then deleting the file. Still no joy.

All the videos I've watched and documents I've read seem to indicate its an immediate reclaim operation. That makes sense because of the zero detect ASICs. Its 3par advantage

The 7400 is running 3.1.2 mu2 p24 (i think that is the patch level).

What gives?

Nick
v800
7400

In my experience, you may have to wait longer than 8 hours. In addition, the Management Console sometimes needs to be re-loaded to refresh the view properly.

I was using both the GUI and the cmd line to validate the process. As of this morning there is no change to the luns. Nothing has been reclaimed and no reduction in VV size.

Any other ideas?

Nick

Do you have the license(s) for all the Thin stuff? Did you set the zero_detect policy on the VV?

setvv -pol zero_detect VOLUMENAME

Yes, I have all the thin licensing bits. In my version of the Inserve OS, zero detect is enabled by default. I've also verified its enabled. I have a v800 frame with esx hosts attached. That frame is running 3.1.2 mu1. I'll see if it works over there. Not sure what I'm missing...if anything.


-Nick

So far as I can tell....it has not worked against the v800 frame either. I may open a case, just curious if anybody has been able to get this working?


-Nick

What happens if you create a NEW TPVV and writes zeroes to it?

The v800 frame started showing space reclaims after a few hours. It is running 3.1.3 mu1. The 7400 frame is past 24 hours now and ot even a single gb recovered. It runs 3.1.2 mu2

-nick

What commands are you running to audit the reclamation process?

I would recommend using:

showvv -s VVNAME

Here you will be able to distinguish between Usr space that is USED vs RESERVED. As your thinning occurs, you should see the USED trickle away every so often... the Reserved space will take some time for a background process to clean up and free that space back to the CPG.

_________________
Richard Siemers
The views and opinions expressed are my own and do not necessarily reflect those of my employer.

I was using showvv to watch the luns. I was using esxtop to watch the process from the esxi side of things. I could see the blocks being zeroed/deleted out from esxtop. and nothing happening onthe arrays.

The v800 has started thinning after 3 hours. The 7400, nothing after 24 hours. Maybe it has to do with i/o across the array and when it has free time?

v800 = 8 nodes @ 1700 drives

7400 = 4 nodes @ 220 drives

-Nick