Diary of a Network Geek

A screensaver that monitors your server.

This may not seem like much of a “tool”, since it’s pretty passive, but when you have a server that’s getting old and failing, being able to quickly glance at what its performance is like can be a real benefit.  And, for the past eight months, I’ve been running a server that, to be honest, was a little too old to be in production.  People would complain about it pretty much constantly, even, I suspect, when it may not have been the actual problem they were suffering from.  I’ve since replaced the server, but I kept running Systemometer on both the old server, which now serves as strictly a backup server, and the new, shiny, Windows 2012 server, just so everyone can see the difference.

But, let me be really specific about this; Systemometer is a configurable monitoring tool that shows system performance and resources in a spider chart. Seeing the varying shapes of displayed polygon, representing a visual pattern drawn based on normal, or critical, system states. Once you get used to it, this snapshot view helps to read the overall system status at a glance. Just looking at the screenshot in this post, which will enlarge if you click on it, you can see that a lot of information is displayed.  Notice, for instance, that there are 12 “CPU”s listed.  Since this is a modern, multi-core server, those are really just all the cores being displayed, with the current processor time in yellow and the average processor time in green.  If I wanted to, I can also set Systemometer to display the maximal processor usage, but as this is a new server, I haven’t bothered to set that.  The same goes for the number of processes the server is handling, as well as the number of threads.  Also displayed is the physical and virtual memory usage, total drive space used and the hard drive seek time.  Notice how almost everything falls well within that red circle on the display?  That’s because the server is new and being used well below it’s capacity, by design.  This is the second server upgrade I’ve done since I’ve been at this company, and I’d like to not have to do one again soon.  That’s also why the number of threads is reading like it’s in the red, even though it’s not.  The new server is so new that not all metrics have been calibrated to display correctly.
Also, notice that the two performance polygons are yellow and green.  The yellow is the current usage while the green is the average usage.  It may be hard to tell the difference between the two because I took this screen shot on a Sunday afternoon with minimal usage.  Of course, the server being primarily a file server and an Active Directory server, the average usage is pretty constant in any case.

It may not be obvious from the screen shot, but I’m running this as a screen saver, which is only one option for using Systemometer.  It can also be used as a kind of replacement performance monitor instead of using the built-in Task Manager for that function.  Actually, one way I validated the results I was seeing from Systemometer was to run it next to Task Manager and compare the displayed performance information.
Personally, I like running it as a screen saver because I can quickly check on my server as I walk past the screen into or out of my office.  Also, it seems to impress people who see it and can’t make heads or tails of what they’re seeing.  It’s not big, fancy monitoring system, but unless you really know what you’re looking at, the average person isn’t going to figure that out!

Finally, the other reason I use Systemometer is that it’s free!
Yep, that’s right, absolutely free.  Of course, it may not ever get updated again, but I’m okay with that, as long as it still works as it has been so far.

I really don’t mind flying with my boss, but I wish he’d stop saying that.

So, Thursday last week, things started to go wrong with our server in Bellechasse, or, as I think of it, the Sweaty Armpit of the Gulf Coast. First, it was a user who just couldn’t seem to connect. Then, there was another user who couldn’t connect, though that turned out to be a totally unrelated problem. After dinking around with the server and the workstation over the phone, we finally rebooted the server and the problems seemed to be solved.

Turns out, not so much. Friday I got a call shortly after 8:30AM letting me know that now four users can’t connect to the server. So, again, after a few minutes of screwing around with a work station, I had them reboot the server, figuring that what worked the day before should work again. Seems like sound logic, right? Well, it is a Windows 2000 server, so logic probably wasn’t the best tool to apply. Everything seemed fine, right up until the server hung up at the “Preparing network connections” message. We rebooted the server at least three times after that and even tried Safe Mode, but, to no avail. So, I broke the news to my boss who was barely able to contain his joy at having an excuse to fly.

I should mention that I’ve flown with him on several occasions without incident and, as far as I can tell, he seems to be a very fine pilot. Certainly the flight itself has been smoother than most commercial flights I’ve been on. And, being able to bypass any sort of security checkpoints or limited schedules is really nice. It means, generally, that I can fly over, fix the problem and fly back in the same day. All with out needing to fill out an expense report, I might add.

So, well before sunup on Tuesday, we flew over to a little flat spot on the edge of New Orleans they call Lakefront Airport. This trip, I noticed that the same roofs seemed to have the blue FEMA tarps over them, but more yards had the pre-distressed FEMA trailers in them. Not sure if that’s progress or not, but, somehow, it seems like it should be. At least more of the traffic lights were working. Certainly, that has to be considered progress.
In any case, we got to the office about 9:00AM and I walked right to the server and got to work. I started by rebooting, just to see if I was going to get lucky. Naturally, I wasn’t or this post would’ve just ended. So, I cycled through several different boot options and finally, after much weeping and wailing and gnashing of teeth, I got the server into a semi-stable state by booting into Safe Mode with Networking Support, but via the Repair Mode on the Windows 2000 install disk. (No, don’t ask me to recreate the steps because I wasn’t taking notes and I was probably running a fever.)

So, guesses on what was wrong? Out of disk space. Or rather, there wasn’t enough disk space for Active Directory to run properly. So, I killed a bunch of temporary files and cranked down on the size of the virtual memory paging file. Sure enough, when I rebooted into “normal” mode, the server came up and everyone was able to log in and all was well with the world. All done before lunch, I might add!
And, so it was that I humbly asked to be brought an oyster po’boy, dressed, sans tomato, with fries for lunch while I attempted to kill all spyware and adware and other such nastiness. What I got was a catfish po’boy, two hours later, and one machine that still has some spyware remnants on it that need to be cleaned up. How fleeting is glory… Oh, and that’s not to mention the several requests I got that were far, far outside the scope of “fixing problems”, which was, in fact, my stated purpose.

Interestingly enough, when we touched down again in Houston an essential piece of navigational hardware in the boss’ plane locked and threw an error message. Then, up popped the infamous Windows NT “Blue Screen of Death”, upside down on that oh, so essential monitor. Yes, boys and girls, our lives depended on hardware that runs the most unstable, buggy, crash-prone version of Windows I’ve ever worked with.
Yeah, we sure cheated death again. Barely.