Upgrade done. It's looking good for distribution of VM on local SSD.
We upgraded he composer and broker to 5.2.
This is a two host local SSD pilot ( for a 50 host project).
The main pool was out of control in 5.1 and after a while only wanted to create VMs on one of the hosts.
Long story short the existing main pool (that can go up to 90 VM) did load balance the VM on the two host after the upgrade. After the View upgrade I changed the pools config for the new image (with patch tuesday and 5.2 agent), I deleted all available and provisioned VMs and let View build back the pool to 80 machines... and they were bablanced properly
The 20 or so connected users stayed connected duing the upgrade without problems.
New pools created after the upgrade also balanced the VM.
Mind you this all worked at the begining of our 5.1 installation so we have to give it some time to be definitive.
One scare that almost sent us back in rollback mode : it looks like the calculations for storage overcommit have changed. During our first provisioning test a few machine got created and then I started getting "not enough disk space for provisioning" type messages when there was actually plenty of space on the SSDs. We googled the message and the vmware kb mentioned that it was either disk space or permisions for the vcenter account. Since we had plenty of disk space we focused on permission (we manage View but not ESX so there was a possibility that it was a permission issue since the vcenter account has limited permission). In the end, as wild guess, I changed the pool storage overcommit to aggressive and all error message disappeared.
"Provisioning error occurred on Pool Desktop_ID because available free disk space is reserved for linked clones"