Sometimes Linux scares me. 452 days of uptime on one of our primary production servers. She just sits there and chugs away processing hundreds of millions of SQL queries and constantly building and rebuilding our source tree looking for errors. Occasionally a process zombies out or the tape drive hiccups, but now it’s a personal vendetta to troubleshoot the system without a reboot. The system has been moved twice…while still attached to the UPS.
But someday that machine will get rebooted. Be it voluntarily or accidentally, a year and a half thick slate of progress will be wiped clean, a new kernel will load, and the infinite loop will start over.