I tell you, I was an HP-UX sysadmin into the late '90s and the regional telco used a LOT of HP-UX.

Around '95 I spent a solid year setting up a pair of T520s worth about a million bucks, to be a HA cluster responsible for part of the billing process, which was being ported to Unix from the IBM mainframe by a team of 20 (mostly inept, a few smart cookies) programmers. Only to be cancelled at literally the last possible moment to keep on the mainframe. I highly suspect that it was all a ploy to get better mainframe upgrade terms.

Not on April 1st, but at one point management spent the last of their budget for the year on upgrading this pair of T520s from 2GB to 4GB of RAM. BUT they didn't buy extra drives to grow swap, and we were already WAY into the deployment so we couldn't just go repartitioning. HP-UX required all memory to be backed by swap to be able to use it, so the extra 2GB of RAM went entirely unused.