Up 401 days!
Sun Sep 26th 2004, 3:39pm
I remember reading a funny story about a group of university sysadmins who "lost" a system on their network. At first, no one noticed, but after a while they realized that it had been running for an insanely long time. As it hummed away in some unknown location, people continued to log into it, jobs continued to run, and while it became something of a support dilemma for the group, it didn't seem to require any physical maintenance, so they just left it in production :-). Turns out, the system had been accidentally dry-walled into a wall, somewhere among the many forgotten corridors of their CS building. They finally stumbled across it one day when someone noticed an ethernet cable running out of a wall.

In the grand scheme of things, uptimes aren't really that big of a deal--these days anybody can set up a Linux box, sit it in a corner, and do nothing with it except watch its uptime creep upwards and upwards. But for some reason it's still exciting to see a vital box stay up for a long time. A combination of factors: hardware reliability, OS stability, good OS security, local power reliability, and maintainer patience (i.e. not Linux-kernel-of-the-day types) all contribute. My primary server at home, sumatra, runs good old, rock-solid FreeBSD. It provides SMTP (Postfix), IMAP (Courier), NFS, NIS, Samba, DNS, syslog, and SSH access for me. In any case, 401 days is a personal record for me, so here it is:
[12:09pm] sumatra:~-> uptime
12:09PM  up 401 days, 17:24, 2 users, load averages: 0.00, 0.00, 0.00
[12:09pm] sumatra:~-> uname -a
FreeBSD sumatra.kehlet.cx 4.7-RC FreeBSD 4.7-RC #2: Sat May 31 20:04:49 PDT 2003    
kehlet@sumatra.kehlet.cx:/z/freebsd/obj/z/freebsd/src/sys/SUMATRA  i386
As you can see from the load of 0.00 it's not terribly busy, but it really is in constant use.

I've been wanting to make some changes to my home network for the last few months, so sumatra will probably get bounced around a bit fairly soon. Either that, or a rolling blackout will get it :-).