The Life of a Sysadmin

Just spent an hour trying to debug why a simple Nagios check script was not working. It basically ran lynx -dump | grep desired string, but for some reason was utterly failing to work.

Eventually I thought to get the script to print out its environment. It turned out that my own environment variables had leaked to the nagios program itself; as a result, lynx was trying (and failing) to open /home/hugh. /etc/init.d/nagios did not (properly? perhaps) clean the environment as I assumed it had. I changed my Makefile to run env -i /etc/init.d/nagios restart, and now it works just fine.

(Incidentally, I love Makefiles as a way of scripting stuff you run over and over and over again. Yeah, they're clumsy and I'm not doing anything I couldn't do with a simple script -- but it's a timesaver to just run "make" and be done with it.)

Okay, I feel like a bit of a tool for never realizing how cool suspend-to-ram is in a laptop. My new laptop for work is a Dell D630, which I'd got 'cos its hardware is pretty much completely compatable w/Linux. However, I've also figured out that a) Ubuntu does suspend-to-ram quite nicely (aside from a couple times when the keyboard doesn't work, but closing/reopening the lid makes it work), and b) it just sips — sips, I tell you! — from the battery.

Now to try and make it work on my own laptop, which is currently sitting at the shop waiting for me to pick it up.

Today's agenda:

• Install new 48-port switch in server room
• Update Fedora Directory Server wiki page on building RPMs for/on CentOS
• Set up mail server to accept mail for older, semi-deprecated domain
• Drink coffee, catch up on sleep

See? I am still a sysadmin.

I just noticed that the Free Software Foundation is putting together what they call a "book sprint" — kind of like the 3-day novel writing contest — to write an intro to the command line for newbies. They're hoping to get it done by next Monday (!).

I like the idea of this project a lot; if I can get some spare time this weekend, I'll definitely be dropping by.

We've got a new server room being built right now; it should be done in about six weeks, so I'm putting together an order for bits and pieces that I'll need.

I've mentioned before that cable management is one thing I get obsessed about, so this site is like porn for me. I'm not shilling for them; haven't ordered from them, no idea if they kill puppies in their spare time or what, but holy CRAP this is all the stuff I've ever wanted: RipWrap (so that's what it's called!), label printers, 87 varieties of zap straps, and I don't know what all.

Wow. Just wow.

Edit: Okay, seriously. There's some really good stuff in here among the advertisements.

With the move to the server room coming up in a couple months, I've been spending some time trying to lay out the racks we'll have there. My current layout is in an OpenOffice spreadsheet; I thought I'd try some other tools and see how they shape up.

• APC Configurator - Windows only. I did try this a while ago and found it wasn't bad, but no power calculations — one of the reasons I went to a spreadsheet in the first place.
• AusrackID - flash based, so works on a Firefox on Linux. Very nice, but a limited number of hardware choices — so there are Apple Xserves but no Thumpers. There's no way to add hardware and no real generic choices ("4U server", "12-disk 8U array", etc). Also no power calculations.
• RackTables - GPL'd LAMP app, so dead easy to install. Not bad at all, but it's an early version (0.16.6) so the interface is a bit clumsy interface and there are lots of features planned RSN. Aims to be a cross between a server room planner and an asset tracker, so that might not fit in with my planned use of GLPI. No power caculations, though there is a request to add SNMP monitoring of APC PDUs.

Still sticking with a spreadsheet for now; it's not the best, but it is flexible and quick. Any other tools I missed?

Actually for a whole office. Excellent reading. Wish I'd known about this at $JOB-2...

NetSNMP uses 32-bit counters for disk sizes. Guess what happens when you've got one of these?

Due to be fixed in the next release, so at least that's something.

...that TCP Offload Engines (TOE) were so detested by Linux kernel folks. The arguments here make interesting reading and seem convincing to me.

(From Andy Grover's blog.)

Thanks to Undeadly and ossowicki for the pointer to wmname, which fixes the grey java windows problem when using Awesome or other tiling window managers. No more starting up Gnome or IceWM to use NetID or Strangebrew, hurrah!

I really dug Charlie Stross' Halting State (link throws the author a few shekels). But now he's declared it obsolete.