HPE Storage Users Group

I have a V400 with the entire thin suite. I have LUNs presented to vCenter 5.1. All vms are eager zeroed and formatted with VMFS 5 and all run thick disks. I do however have 3par Thin Suite running on the 3par V400 array which keeps the LUN truly thin. As we of course move around vms between datastores, etc the LUN space used grows in 3par Management Console and the percentage keeps getting higher.

I had read about re-claiming space on LUNs and zeroing it out. I am to use vmkfstool -y 60-99% in order to reclaim this space and zero it out.

I can run the command without issue, watch it grow and then delete the balloon file without issue. It however never changes the percentage used in 3par Management Console? I am guessing I am missing something extremely simple?? Any ideas what I am screwing up/misunderstanding?

You're probably looking at "Reserved Space" rather than "Used Space" see this thread :-
http://www.3parug.com/viewtopic.php?f=18&t=516

Also take a look at the TP whitepaper
http://h20195.www2.hp.com/v2/GetPDF.aspx%2F4AA3-8987ENW.pdf

Thank you for the information! I am still digging into it, but pardon my ignorance as I am still confused?

I have zero detect on all TP LUNs, I have moved vms off of a datastore and it isnt even half full, yet in Management Console for the TPVV, it shows: Reserved User Size (% Virtual) as 91%.

I have tried in the doc you linked to do a vmkfstools -y AND the -c commands, still nothing. All the vms are eager zeroed, so I shouldnt need to run sdelete or dd within the OS? What am I missing? A tunesys/tunevv/compactCPG do not help in this case....

How long have you waited? It’s not a real-time process

I read about the garbage jobs and the other job that can take 8+ hours, it has sat overnight and honestly been this way for 12+ hours now.

You shuld have seen some reclaim by now. Are you sure that the view is refreshed?

look at used user size, you can add this into the GUI by right clicking on the reserved user size column header. As mentioned earlier the garbage collection will do the least work possible assuming ongoing growth is happening in the CPG.

It finally started going down. I was refreshing the client and even re-opening it so I am not sure what was going on. Thank you for all the help everyone!

Hi.

> As we of course move around vms between datastores

Why do you move VMs between datastores on regular basis?
What's the point?

Yizhar

We are constantly rebuilding client OS vms, migrating groups of vms to load-balance the I/O requests, etc. We boot up hundreds of clients at a time depending on the test case and have hundreds of server vms running. As we never know what client vms are needed and where, we can end up with saturated and heavily hit datastores.