I worked at a company in the late 90’s that transitioned from Netware ( Novell ) to Microsoft. The company was growing like crazy and always renovating.
At one point, all the NetWare servers had been retired but one. They could not find it. Eventually it stopped being used but it was still on the network for a couple of years. One day, they tore down a wall to make an open area and there it was! It had been locked into an inaccessible empty space between walls but just kept running.
I never knew Novell myself ( other than the bulky adapters my laptop needed to get on the network ) but that server made a big impression on me. Legendary.
We had one of those at the old AMD campus in Austin!
I was one of the lackeys that had to go out and put hands on a list of machines we couldn't physically locate (engineers had moved them around the office for various reasons and we'd lost track of their current home). After several days, the list of ~100 machines was whittled down to just one. We knew which network port it was on but couldn't physically locate it. We had the speaker playing row row row your boat and walked around listening closely for it (even after hours) but never found it. We knew it was __ft down the network cable, but that was it - no idea where that cable actually went.
Finally we lifted ceiling tiles and traced that little cable to the top of a partition/wall... the cable dropped down into the wall with an extension cord. In the darkness we heard it dutifully still playing row row row your boat. We peeked over the wall with a flashlight and there it sat.
shutdown was issued, the extension cable was pulled but it pulled free and the poor, retired, powered down workstation clunked into the bottom of its void. We noted its odd location in the wiki and called the whole project complete.
It may well still be there even though AMD has moved away across town.
The old server-that-got-drywalled-in meme was something we as grad students in the late '90s talked about doing intentionally, to keep Internet-accessible research prototypes running after graduating.
That service or even tech report archive coming from a .edu research group wasn't in a machine room -- it was under someone's desk, or out in the lab area of old junk computers used by undergrad assistants.
We had the support of a PI. But some schools select for people who are focused on their own success. So, if unplugging an unknown machine removed a barrier to someone's immediate goal, when we were no longer there to guard it, we couldn't assume they'd pause and think, "Maybe I should contact this email/phone on the big laminated sign on this machine, before getting it out of my way."
I recall that discussion ending with the idea of the drop ceiling being the most likely place to keep it undisturbed. The exposed cable run trays for fiber and such were already right up by there. But I don't think any students from our group ever did that.
Later, the tragedy around Aaron Swartz included him stashing a machine much like that.
It's hard to explain to outsiders how MIT culture (including down from the top) celebrated having few rules, taking the initiative, and being open to a community of loosely/flexibly-affiliated people. The official MIT Museum even celebrated mischief that was possibly illegal.
I don't know whether the outside authorities understood that Swartz's activity was closer to normal/approved than it would seem to an outsider. Some activity had to stop, but I think the culture would've been to nudge the curious and energetic young person onto a better path.
At one point, all the NetWare servers had been retired but one. They could not find it. Eventually it stopped being used but it was still on the network for a couple of years. One day, they tore down a wall to make an open area and there it was! It had been locked into an inaccessible empty space between walls but just kept running.
I never knew Novell myself ( other than the bulky adapters my laptop needed to get on the network ) but that server made a big impression on me. Legendary.