Wednesday, June 11, 2014

Provisioned Space and Used Space show 0KB

Ran into this issue earlier today when I was trying to get sizes of VMs in vCenter. A half of my machines were showing 0.00 B for their Provisioned and Used Space.

The solution was simple enough after scratching my head for a little bit. Select the server in vCenter and go to Configuration Tab. Select the Storage Adapters under Hardware. In the top right hand corner you will see Rescan All... click that and click OK at the next screen. vCenter will now fires off a rescan of your HBAs and your VMFS volumes and does a "Recompute Datastores Groups". Once that is done it should now show your Provisioned and Used Space without issue.

Sometimes it is the simple things that drive you crazy!!


So I was running through the patching schedule with one of our other admins, when I decided to opened up RVTools and started looking at the vHealth tab. I was shocked to see :

As hard as we try to keep our environment clean, this was unexpected. What in the heck is a Zombie file? What are they doing? What the heck is causing them, and how can I get rid of them?!

Turns out a we had several different kinds of Zombie files. Our first kind was in our normal vSphere 5 Server environment, a snapshot that was not being shown in snapshot manager. Bizarre right? What they were doing is sucking our brains! Eh, ok so it was only eating up storage space on our LUNs, because they were not being cleaned off, and were causing the VMDK files to grow out of control. Same difference right? The second Zombie files that we had were being caused by our Horizon View VDI environment. These were different. These were old copies of our Virtual Desktops that hung around after we did recomposes.

The cause of the first set of Zombie files was a leftover snapshot from our EMC Avamar. Still investigating what happened here as this is not normal behavior but we do know it did it to a bunch of our VMs. The fix is fairly easy although it is tedious. Go to the VM take a snapshot, sometimes the new snapshot will fail but once it has failed it shows the old snapshot in the Snapshot Manager, other times it will succeed and you will now see two snapshots in the Snapshot Manager. Either way the solution is the same, select the Delete All in Snapshot Manager and wait for the snapshots to clear out.

The second set of Zombies, the VDI Zombies, were a little different. We had to go through the DataStores that we have assigned to Horizon View and check the dates on the files. Since we are using Dynamic Pools our VMs are recreated everyday, so we knew that if the files in the folder did not have today's date on them we could delete them. This was caused by creating "new" images. What we actually did was cloned our old golden image in an attempt to flatten out how many snapshots we had on it. Since it is technically a new Golden Image, the old folders that had the VDIs stayed out there, and the new ones were created as well. Again, another storage waste!

Now I know this probably all sounds very novice, but as I couldn't find anything on this with a "quick" 20 second look on the web so I figured I would share. Better policing of our DataStores is obviously needed! We are getting ready to implement (or at least demo) vCOps. Hopefully that will give us some more insight into these things, but still a HUGE shout out to RVTools !!! I love that tool! Great place to see little things in your environment that might not be right. It certainly saved my bacon!


VMWorld 2014!!

I am registered !!! Looking forward to seeing everything that has changed and being immersed again in all that is VMware!

I think I am going to take a crack at my VCP5-DCV while I am there. I haven't cert'd in a while so while I still can without having to take another class I guess I will get to it!

Hope to see you all there!