It’s no great surprise that my experience with Horizon (I’m of the age where I remember when it was View) in recent years has me rather excited to see the General Availability (GA) of support for Nutanix AHV.
This has been a long time coming with Omnissa spinning out of VMware into its own entity with the Broadcom acquisition.
I really believe this finally gives Horizon the opportunity it really needs to seen as a vendor agnostic broker platform, rather than just ‘VMware’s EUC solution’ that often meant a dependency upon ESXi. For many years however, we’ve deployed solutions using physical capacity such as HPE Moonshot (RIP), or via manual desktop provision on other hypervisor solution – but it’s a fair statement that ‘most’ Horizon solutions depended upon a VMware vSphere platform for the desktop / application hosting.
Fresh under the Omnissa re-branding, after a very swift Beta process a minted release Horizon 8 (2512) landed, to much excitement here.
It didn’t take me long to spin up a quick Horizon Connection Server natively in AHV, add in a Capacity Provider (connects to Prism Central, for multicluster magic) and spin up an automated non-persistent pool for testing.
For those experienced in using Horizon, the creation of a desktop pool is virtually identical to what you’d normally see, including the use of Cloneprep (no more sysprep delays).
I did find a tiny little bug when deleting a desktop from a Capacity Provider (Nutanix) under the ‘Machines’ menu it incorrectly states Amazon WorkSpaces. (under the Desktops menu is correctly says Nutanix).
Having a little dig around I found that Horizon creates a category in Prism Central and places VMs in either ‘InternalVMs’ or ‘PoolVMs’ which will help with addressing a common frustration experienced vSphere Admins have in that there’s no VMs & Templates view with folder hierarchy in Nutanix.
I do need to dig a bit further into this, I created a single pool with 25 VMs, I expected a single InternalVM and 25x PooledVMs but seem to have got 3 and 26.