Clustering, clustering, clustering...
Ok, today I presented at LILUG about clustering. Unfortunatelly, the whole talk sucked! Here’s what happened….
- I had about 66% of the talk prepared, which was fine considering that I didn’t have to talk about _everything_ I planned, but…
- I decided to make few last minute changes, which worked at home. They caused me to be a little late (15 mins, that’s not too bad, I didn’t intend to start until 30 mins into the meeting anyway..)
- I got there, set up the network, etc., etc.
- I boot my laptop, everything goes fine, except the projector & Adobe Acrobat Reader
- I tried making the resolution useable, but acroread insisted on using the secondary display to do the full screen mode.
- 15 minutes later, I gave up on it, and tried the KDE PDF viewer, that worked, great.
- I boot my desktop which served as a DHCP/DNS/TFTP/NFS server
- I tried to ssh into the desktop - no luck
- Checking the network cables yields nothing
- I hook up the desktop to the projector and reboot - aha! Lilo doesn’t work
- I put in a debian install cd, get a shell, chroot, and run lilo
- reboot, it works
- I connect the projector to the laptop again, and I’m ready to start, this is 1 hour after the meeting started, and there is 1 hour left.
- I go through the slides, which were really great looking thanks to LaTeX and D. E. Knuth
- I finish with the slides in 20 mins, and I want to demo something, but the technology failed me agian…
- I just take 40 mins of Q & A
I guess you’d call it a success, since I think I had about 8/50 people understanding at the end of the talk. Someone also noted that clusters are fragile enough that mostly they barely work if you don’t touch them, if you do (like moving it elsewhere) they tend to break big time.
Anyway, that’s it for today…I’ll send the mailing list some of the materials tomorrow, and we’ll see.
I should add that as always, I tried to present about too much at the same time :-)