<font face="Verdana, Arial, Helvetica, sans-serif" size="2">Originally posted by ClueByFour:
I guess the point I'm trying to make is that Linux is far from an end-all solution to all things. Like many things, it has it's place, and the trick is to know when to deploy it, but with the proper expectations.
Although I'm getting further and further away from the hands-on side of the business, it strikes me that uptime of a particular box is nothing to write home about--in fact, uptime of a year suggests to me that the admin has been rather lax about security and performance patch application (although clearly this depends upon the environment in which the box sits).
One of the nice things about Linux (and why the stuff I've run on it in the past never had silly "uptime wars") is that I can afford to get me a nice cluster of n + 1/2/x machines, such that I can pull them offline to patch without disrupting the app and service that is being provided.
The short version: if the box has an uptime of a year, and has been rooted for 300 days, is the uptime really that impressive 
?
</font>
Virtually all linux pathches and upgrades can be installed without rebooting the entire machine. As a matter of fact, the only time you do need to reboot is if you're installing a new kernel. Most kernel exploits require local access to the box and are beyond the realm of the script-kiddies.