Diagnosing a Server Crash
A repeatable method for finding what killed a server, starting with the three things that cause most crashes and are trivial to rule out.
Everything published under this tag, newest first.
A repeatable method for finding what killed a server, starting with the three things that cause most crashes and are trivial to rule out.
What still works with no connection, what quietly does not, and how to set up a game for a room full of people.
Choosing a mode is a decision about where there is room, not what you like. How to spot the gaps, and how much moderation each mode quietly costs.
Splitting a population is usually a mistake and occasionally the only fix. The arithmetic that decides it, and what it costs in moderation.
Three tools with genuinely different trade-offs around cost, control and latency, and which kind of community each one actually suits.
Screen, tmux and systemd solve the same problem differently. Which to use, and why the answer changes once the server matters.
The dozen commands that cover almost everything a server owner needs, and the four mistakes that cost people their machines.
Copying the files is easy. Keeping the address, the players and the search ranking is the work, and it starts a week early.
The update that takes a server down is almost never the game's. A staging habit that costs twenty minutes and saves a weekend.
Left 4 Dead 2 still has dedicated community servers. Campaigns, versus, custom maps, how to join one by IP and what to check before you commit an evening.
On Goldrush one team escorts a tank up a hill to blow open a bank, then carries two crates of gold down to a truck. It takes twelve minutes when it works.
The first troll that finds your house does not open the door. It swings a tree at the wall, the wall stops being a wall, and then the wooden roof falls in.