While doing some surfing the other night, I came across a really neat site called the “Uptime Project”, where members “compete” by basically seeing who can keep their servers running the longest without a reboot.

From their web site:

The Uptime-Project is a Fun-Project. Using our service you can collect all your uptime-data (your computer’s uptime) and compete against other users.

The gist is simple. You sign up, then download and install a tiny client on your server. The client software (through “cron” or whatever scheduler you use) occasionally pings the competition server with your current “uptime” statistics (i.e. how long it has been since you last rebooted). Currently, the client is available for Windows, Linux, BSD, and SunOS.

I decided to sign up for fun with one of our custom-built (dual Xeon) Linux development servers. We debuted within the top 50 (out of nearly 12,000 registered servers). Here is a screen shot (via SSH) of our current uptime on that box:

495 days, or, 1 year and 130 days. Not bad! At least, I thought so, until I saw the guy sitting at number one. His SunOS box has been running non-stop for over six (6) years! Madness!

The downside to all that uptime, though, are the obvious security issues … surely at some point, you’d need to upgrade the kernel and do a reboot. The top guy is basically running a six year old operating system. That is the equivalent of running Windows ‘98 on your desktop here in 2006.

Check ‘em out at: www.uptime-project.net

BTW, that hostname is “rampage” - a nod to my old Army unit (3d Battalion, 64th Armored Regiment … the Rampage battalion!) If you are curious, the specs on Rampage are:

• Custom built 2U rackmounted server
• Dual Intel Xeon 3GhZ processors (with hyperthreading - 4 virtual CPUs)
• 4GB RAM
• 15K rpm SCSI RAID array
• Running Linux 2.6.xx SMP

Cheers.