Has anyone had issues with their instant clone VDI pools where users begin their connection to an instant clone and it does not connect over PCoIP but will connect to the same system over VMware Blast. Once the user logs on, you can disconnect from the system and reconnect to it over PCoIP.
User 1 > PCoIP > System 1 = FAIL, blank screen and automatic disconnect
User 1 > VMware Blast > System 1 = Connects, user then disconnect from System
User 1 > PCoIP > System 1 = Connects
This is happening for the same systems in the environment, perhaps 5 or 10 systems out of 150 keep doing this. The system names are recycled, so System 1 will cause issues for the User 1 and while they are fighting that issue User 2 logs on and gets System X and has no issue using PCoIP. The next day User 2 logs on first and has issues connecting to System 1 and User 1 connects just fine to System X.
I am getting ready to recreate the instant clone pool with a new naming scheme and would like to know what is causing this strange behavior. DNS and DHCP entries match up properly and it is always the same few systems that are causing issues.
Update: I kept logging in and out of systems until I reached one that displayed the behavior. Once identified I went to vCenter and selected the system. The preview window showed that the desktop logon screen was squeezed, as if three monitors had been in use before and it was stuck on that setting. I opened the remote console for it and after a bit of waiting it finally resized to fit the remote console window. I then re-initiated the Horizon connection and it connected to the system over PCoIP. Perhaps PCoIP is not able to resize the windows properly and VMware blast is able to fix that. My parent VM has 42MB of video but the instant clone pool allows for up to 3 monitors at 1440p. Does Horizon modify the base VDIs allocated video memory or does that have to be done on the parent VM?
