4 ms·

I figured HN probably ran on an old desktop tower under somebody's desk, with a post-it note on it that says "DO NOT REBOOT!!!!"

oefrha 5y ago
In fact, restarting the server is or was the only way to fix at least one bug on HN.[1] Pretty sure I've seen it as late as last year, and dang confirmed the fix was still the same. (I know, restarting the server != rebooting the machine.) [1] https://news.ycombinator.com/item?id=23228692 https://news.ycombinator.com/item?id=23228692
pvg 5y ago
https://news.ycombinator.com/item?id=23472870 https://news.ycombinator.com/item?id=23472870
- acheron 5y ago
  Lol, yes. Maybe we’ve worked at the same places.
bierjunge 5y ago
This reminds me of my old workplace. Government owned company where everything was running on MS Excel, SAP, fax machines [0] and prayers from operative IT office. I had to go to one of the more "exotic" server rooms, placed in a basement and saw a very old desktop stuffed on the bottom of a rack. Covered with dust, open side panels, plastic turned yellow, you know how old and forgotten hardware looks like. This had a new-ish post-it with the message: "this runs all of the fax infrastructure. you turn it off, you explain it to the CEO". [0] https://en.wikipedia.org/wiki/Fax https://en.wikipedia.org/wiki/Fax
- robbedpeter 5y ago
  I worked for a company that had acquired a B-list, top 20 mmorpg - its database was mssql and had been running 24/7 for 5 years, with no way to back it up, no safety net, and an increasingly pissed off player community suffering degrading service. Everyone was terrified to reboot. They started bouncing paychecks before I got to see the drama come to a head, but I'm obsessive about proper levels of redundancy anymore. Don't build fragility into a system, and at least leave yourself a means to fail gracefully.

4 ms·

I figured HN probably ran on an old desktop tower under somebody's desk, with a post-it note on it that says "DO NOT REBOOT!!!!"

by acheron 5y ago

oefrha 5y ago
In fact, restarting the server is or was the only way to fix at least one bug on HN.[1] Pretty sure I've seen it as late as last year, and dang confirmed the fix was still the same. (I know, restarting the server != rebooting the machine.) [1] https://news.ycombinator.com/item?id=23228692 https://news.ycombinator.com/item?id=23228692
pvg 5y ago
https://news.ycombinator.com/item?id=23472870 https://news.ycombinator.com/item?id=23472870
- acheron 5y ago
  Lol, yes. Maybe we’ve worked at the same places.
bierjunge 5y ago
This reminds me of my old workplace. Government owned company where everything was running on MS Excel, SAP, fax machines [0] and prayers from operative IT office. I had to go to one of the more "exotic" server rooms, placed in a basement and saw a very old desktop stuffed on the bottom of a rack. Covered with dust, open side panels, plastic turned yellow, you know how old and forgotten hardware looks like. This had a new-ish post-it with the message: "this runs all of the fax infrastructure. you turn it off, you explain it to the CEO". [0] https://en.wikipedia.org/wiki/Fax https://en.wikipedia.org/wiki/Fax
- robbedpeter 5y ago
  I worked for a company that had acquired a B-list, top 20 mmorpg - its database was mssql and had been running 24/7 for 5 years, with no way to back it up, no safety net, and an increasingly pissed off player community suffering degrading service. Everyone was terrified to reboot. They started bouncing paychecks before I got to see the drama come to a head, but I'm obsessive about proper levels of redundancy anymore. Don't build fragility into a system, and at least leave yourself a means to fail gracefully.