Mastodon Feed
Reblogged by jsonstein@masto.deoan.org ("Jeff Sonstein"):
lauren@mastodon.laurenweinstein.org ("Lauren Weinstein") wrote:
Once upon a time -- many, many years ago and late one night -- I entered the ARPANET (the Internet's ancestor) computer room at UCLA to do some work on ARPANET host #1 -- the DEC PDP-11/45 that was UNIX UCLA-ATS.
But something was wrong. The system was behaving very oddly. It was up, and not under heavy load, but ... it was as if it kept starting and stopping. When I'd type there'd be activity -- I could see it on the front panel lights, but when I stopped typing ... everything seemed to stop. I could get keyboard echos, and then a response ... but then everything would stop again.
Normal boot wasn't working either, but it was possible to reboot from the emergency DECtape.
But still clearly failing overall.
When we ultimately found out what the problem was, I was quite amused. The real-time clock hardware had failed, so the process scheduler was not cycling properly. However, whenever someone typed, the keyboard interrupt would push the scheduler so there'd be some action -- but then it would stop again until there was more typing and more interrupts.
Basically, the system was operating as if on a hand crank. Live long enough, and you see almost everything. -L