Limping Along
Wednesday, April 2nd, 2003Work is quiet for the most part but I've had an interesting chat with one of the Unix SysAdmins on one of the several Linux boxes I have in production at the time. I can't have root access, but they allow sudo access, and that's good enough for me as I can do all I need on the boxes assuming that the sudo command is working. Well... late yesterday sudo started rejecting my password on one of my production boxes. No big deal - I'm still able to login to the box with SSH, but it's something I wanted fixed. Maybe it was in NIS, maybe not, but I wanted them to restart services to see if it was NIS. Big Mistake.
They looked at the box and couldn't figure out the problem. This is not a big surprise as these are not the very best admins I've ever encountered in my days and from talking to others around here, this is not an uncommon belief. Their 'solution' to the problem was to first upgrade the machine to '011' - their numbering scheme for semi-custom RedHat builds, and the second was to say that Linux really isn't stable enough to allow for the 201 days of uptime that this box had.
Clowns.
I told them just to reboot the box as that will most likely clear things up and leave it at that. They think it's important to move to their "latest" release when they don't have a good idea of the problem. Holy Cow! I know it's not a really high paying job, but given that these people have a lot of power, it's amazing that they don't screen them a little bit better than they currently do. Yikes!
OK... first, I can't imagine why on earth they make a custom build of the kernel as well as custom OpenSSH packages (to name just two). These can't possibly be kept as up-to-date as the developers of these same packages, but they insist that it's for security reasons. Amazing. Really.
Then when I have a reasonably stable system (on average > 200 days uptime) they want to rebuild it with yet another of their crazy distributions as opposed to just learning the systems they have and be able to fix them. I know there are some in the Linux Admins that are very knowledgeable and could probably figure out this problem, but they are not available to do this work because they are too busy doing hardware verification, creating new builds of the kernel (I still can't get over that), etc. It's amazing that a place like this is run this way.
But life goes on... slowly, and with great gnashing of teeth, but it goes on.
I read an interesting article today about the role of the Artisan in the human experience. Very interesting thoughts - lots that hit very close to home for me. It's a battle with frustration to be sitting in this place where so much can be done, and yet because of some (understandably) important decisions nothing is being done. I understand business decisions - I owned a business for 5 years, there should be the realization that the business of business is to keep moving forward on things. Never sit still. Unfortunately, that's not being done here. So we limp along.
Maybe I'll plan for what should be done and then see if it's worth implementing this. It'll certainly give me something to do while they make the decisions that will decide what direction we're going to take.