A while back I performed an in-place upgrade on a client's information management system. They were a not-for-profit and had invested little in IT; consequently the "server" was a random desktop host A with an at-capacity HDD, and file store itself physically stored on another hard drive in host B, UNC accessed in the software. Host B's HDD was also close to full.
Upon completing the upgrade, I wrote up a report along with nice drawings showing their system architecture, and a strongly worded caution that the file store was NOT on the "server" and to make sure host B was always switched on and DO NOT TOUCH THE FILE STORE FOLDER.
Whilst on call over xmas some time later, I received a panicked phone call from the client. He'd run out of space on the server, needed to store some things and found another computer with a big hard drive they decided to UNC map to...once he deleted some "file store" folder which obviously wasn't important, so there'd be enough space to use. You can imagine why he'd called our emergency number.
I gave him the bad news and suggested he go to his backups, after which I would assist in fixing the app side. The good news was, he DID have backups, it was xmas and nobody was there except him so there'd been no changes in several days, and he managed to recover everything. I then go to work fixing a bunch of in-app links - scriptable - and went back to watching the cricket while thinking about my on-call pay ticking over as the script ran.
I don't know if he was ever required to explain the bill my company would have charged (or, more likely, a big chunk of time deleted from our balance of paid-for-in-advance support hours).