# uptime
7:02pm  up 443 days  3:47,  2 users,  load average: 0.01, 0.16, 0.50

Finally I had to reboot to do an OS upgrade. This was a SLES 9 box, 2GB ram, 2 dual core Xeons, 5 72GB SCSI disks. It has 25 users logged on every day, network boots 6 diskless workstations, mail, DNS, samba, and misc other stuff. No leaking memory, no borked filesystems; it just ran.