I'm almost ashamed of posting this:
Code:
$ ruptime
erica down 4:08
pc13 up 3:29, 1 user, load 0.00, 0.00, 0.00
pc16 up 13+14:35, 0 users, load 0.08, 0.07, 0.05
pc17 up 8+08:10, 0 users, load 0.00, 0.00, 0.00
pc5 up 10+11:01, 0 users, load 0.00, 0.01, 0.00
pc7 up 1+11:43, 2 users, load 0.32, 0.20, 0.17
pc7 is my computer. pc5, pc16 and pc17 are my servers. The others are my sisters' computers.
I feel I have to make excuses for this.
I had to reboot my system yesterday because of a stupid mistake at the end of one of the cooler things I've done. I had just patched glibc, manually and binary using dd to replace the operand to some instructions after having found and fixed a bug in getaddrinfo(3) which sorted IPv6 addresses erroneously (and yes, I
reported it). Then, when I was to install the patched version, I ran "cp libc-2.3.4.so.new libc-2.3.4.so"
without first removing the old libc file, which truncated the old glibc to zero bytes and caused every single process on the system to crash. Doh!

I had to boot a LiveCD to fix it. I usually get around 3 weeks on this computer before I make some change in the kernel and have to restart it... it's really a testbed computer.
By the way, the FC3-test3 LiveCD kernel hangs when trying to write to a ReiserFS filesystem. I spent an hour at that, so don't fall for the same thing.
As for the servers, it's frigging mad to have so little uptime on them. I used to have over 200 days on each of them. Then a hard drive failed (temporarily... phew) in the NFS server (pc5), I accidently unplugged the power cord to pc16, and the kernel (a FC2 stock 2.6.8 ) failed in the router (pc17), which was really strange. It was overflowing in some routing table cache and refused to flush it, so I had to reboot it. I'm replacing it with a vanilla kernel right now. I'm also planning to replace the system on them with Gentoo in a near future, so I guess I'lll be losing all the uptime on them. Not that it's much of a loss right now... :-/
We've had a few brown-outs over here, but the power supplies in the servers have been good enough to overcome them. I've just bought myself a used UPS just to be on the safe side, though.