In fact, even if it's broken, please don't fix it. It's probably not actually broken; most likely, you've forgotten to plug it in, or turn it on. Or, perhaps you've forgotten your password. Or, you put your peanut butter sandwich into the manual feed tray.
Although those things might sound far-fetched, they honestly do describe some of the things a system administrator faces on a daily basis—at least, those system administrators who deal with tech support regularly. Apart from PEBKAC errors, as sysadmins, we do have duties that require us to accomplish more things than a human being can possibly accomplish—that's where Linux comes in. Not only do Linux servers (in my experience) require less maintenance, but the tools available for Linux administrators are amazing, and the community is wonderful.
Before you solve a computer problem, it's important to find what is actually going wrong. Kyle Rankin gives us the first part of a series on troubleshooting. This month, he helps us figure out abnormally high server loads. Perhaps I shouldn't run Seti-at-Home on Kyle's servers anymore; he's bound to figure out what process is eating up his cycles! As strange as it sounds, occasionally it's simpler to re-install than it is to troubleshoot. Dirk Elmendorf shows us a wide variety of tools to make installation painless and possibly even fun. Speaking of fun, anyone that has been subject to configuring Sendmail over the years certainly will appreciate Mitch Frazier's review of Axigen, an e-mail server with a GUI interface by default.
Nothing, however, makes the life of a system administrator easier than good planning. Jason Allen describes multiple aspects of planning a successful server infrastructure. In many ways, I wish he'd have sent me the article years ago, but even if your server farm is well established, his tips can help turn a nightmare into a dream job, or at least make the nightmare a little less scary.
Once your servers and workstations are set up, security is extremely important. Contrary to many user's opinions, security is not in place to hinder a user's abilities, but rather to protect the user from harm. Yes, sometimes that means protecting users from themselves, but it also means monitoring for strange activity and keeping a consistent interface. Jeramiah Bowling demonstrates one tool that helps sysadmins with that task. AlienVault is a security information management system that provides a common interface for several aspects of security management. If you manage computers, you manage security. You'll likely want to check it out.
Another downside of being in charge of system administration is that computers generally work 24/7. That means we have to be available at any time, and from anywhere. Eric Pearce understands that need and shows off how he gets Nagios to alert him via SMS messaging. Unfortunately, being alerted is only half the solution. If you're in an unfamiliar network, or even a network you know is unsafe, der.hans' approach to SSH tunneling can get you back to your network safely. Sadly, we don't have an article on how to explain to your date why you need to set your laptop up in the middle of dinner. You're on your own for that one.
Now that cloud computing is all the rage, it's possible your “server room” doesn't even exist anywhere other than in some mystical on-line space. Bill Childers demonstrates using Ubuntu 9.10 in Amazon's Elastic Cloud. Feel free to argue amongst yourselves whether cloud computing is the future of server infrastructure or just an annoying fad. But, if you really want some fun arguing, Bill is the one to listen to. No, not about cloud computing, but rather about /opt versus /usr/local. Kyle and Bill, as usual, have drastically differing views on the topic in the Point/Counterpoint column. I'm not sure who I agree with this month. You can decide on your own.