Once again I’m at the office, this time at the tail end of my monthly Traffic server backup. It takes five hours to fill the first tape and about an hour and a half to fill the second… give or take.

I managed to amuse myself, however, by successfully getting City of Heroes to work on the future “lobby video” computer. Hey, it was that or six solid hours of anime and Neopets.

Hmm, wait…

By 9:30 this morning I knew something was horribly wrong. Several computers located all over the building were unable to sign onto the fileserver. The error message, when input to Google and Novell’s support site, turned out to be singularly unhelpful.

By 10:30 I’d exhausted most of the seemingly-obvious solutions. By 12:30 I’d rebooted the fileserver, to no avail. By 1:30 I’d downloaded updates to every bit of software I could think of.

At 2:30 I noticed that someone’s computer was plugged into a “new” jack in that particular office. Oh, hell. The “new” jack meant that their computer wasn’t on the “old” set of racks in the server room, but instead on the “expansion” racks, which meant that…

Yep. One of the network switches in the expansion racks is hosed. Moving the affected individuals’ network connections to other switches fixed their problems completely. If I’d simply traced the damned network connections in the first place I could have saved at least five hours of work.

Of course, actually tracing out everyone’s connections and getting them moved took another hour.

So if you’ve been wondering, “What kind of idiot is this guy, anyway?” Now you know.

I love the fact that someone can save a file in PowerPoint, end up with a corrupted file (at least, corrupted enough for PowerPoint itself to throw up a cryptic, useless error message when someone tries to open said file), and I can come along with OpenOffice.org to save the day by opening the supposedly-corrupt file and saving it to a new, working copy.

…from my Linux box.

Sure, some of the font and formatting information is gone, but at least the AE hasn’t lost the fruits of their labor. Heck, it’s not even their fault… I mean, they are “saving often” like we tell them to, after all. It’s PowerPoint itself that seems to have issues.

Go, OpenOffice.org!

The beeping started at a bit after 2:30 this morning.

“Logger 2 is DOWN!” read the message on my phone.

“Okay,” I said, and crawled back into bed.

Three minutes later, “Logger 2 is UP!”

“Great,” I said, and crawled back into bed.

Thus began an ordeal that continued until, oddly enough, just about the time I left for work. Every twenty to forty minutes, I’d get a false alarm on either Logger 1 or 2. As you might imagine, after a couple of hours I was losing patience, not to mention any chance at a restful night’s sleep. And in case you’re thinking, “Why didn’t you just turn off the phone?” all I can say is, “And that’d be the exact moment that a real emergency reared its head, and if nobody can reach me during one of those then my head could roll.”

I found the problem, though it wasn’t (specifically) with the Logger computers themselves. Oh, no. It was Servers Alive, which I’d reinstalled recently. The detection timeouts for Loggers 1 and 2 were set to 5 seconds, while Logger 3’s was set to 20 seconds. No wonder I never got a false alarm for #3, eh? What’s more, for some reason Servers Alive had reverted to a check schedule of every three minutes during off-peak weekday hours! Hello? Overkill much? Even worse, the weekend on- and off-peak check intervals were set to every single minute. Ouch.

The timeouts are now set correctly, as are the check intervals. The fixes come too late to get my lost night of sleep back, but at least if my phone beeps at me tonight it probably won’t be a false alarm… I hope.

At 5:20 I was just about to grab my bag and go, but was stopped by one of the production guys across the hall complaining of trouble with their GP computer. As I was asking him for details, the guy in the production room next door also complained about the machine in his room… with the same symptoms.

Oh, joy.

At first it looked like a server problem, only those two computers were the only ones experiencing problems. (One of the machines was causing a problem on the server, but once I figured that out I was able to remedy it fairly quickly… albeit not without some frustration.) Full comprehension of the problem came when I realized that those two specific machines are identical vintage-and-model IBM desktop computers, the only two of their kind in the building. Coincidence? I think not!

Now, here’s how this is all my fault in the first place. I rebuilt my software update server a while back, and since the new version of Microsoft’s product allowed me to add “device drivers” to the list of things I could offer, I went ahead and told it to grab available drivers and install them for machines that have the Intel graphics chipsets.

I’ll bet you can already guess which are the only two computers in the building that have Intel 865G video hardware, huh?

Repairing the damage involved hooking up my handy-dandy USB “swiss-army-drive” (loaded with Intel’s motherboard and networking drivers) to both computers, reinstalling both drivers on both machines, and rebooting copiously. Oh, and I told my updater server to never again fiddle with device drivers…

Now I just need to wait for the 7:08pm bus so I can go home at last. Whee!