01.jan.2004
I'm still trying to understand how to use Wiki. I'm seem to be using it as if I were writing web pages. Yesterday I found out about RecentChanges, which explains why everyone could see the pages I thought I was still preparing in secret.
I've spent the last week looking at other Live CD distributions (http://freaknet.org/martin/dynebolic/livecd.html) and the winner for booting speed is... Linux Bootable Business Cards! Second place goes to... dyne:bolic! Then come the others: Slackware, Mandrake, Gentoo, Knoppix, Morphix and all, and the slowest of them all is... Ananda Margii Global Linux Group Spiritual Path Tantra and Yoga Live CD! Maybe these spiritual people have more patience than us!
Oh, and just for the record, OpenMosix is still sitting there poised but unconfigured. *sigh*. I've played a lot of Descent and Doom though, and the Morphix Combined Gamer is a doozie - well, the bits of it that work, that is. Like Dynebolic, when it's good it's very, very good, but half the things are broken and overall it's a mess, but well worth the trip!
10.dec.2003
My plan for today was to get OpenMosix working between two of my own machines. I spent hours fiddling with the power supplies and hardware of nekon2, updated the duplicate of nekon's Unix system onto nekon2 via rsync This has the unfortunate side effect that when it replaces /vmlinuz the machine won't boot any more. rdist can handle this as you can tell it to run commands when specified files are updated, but I haven't figured out how to make it work over ssh yet...
Lots of kernel compilation: recreating the dynebolic kernel from fresh sources plus squashfs and openMosix patches plus jaromil's kernel config. Unfortunately the two sets of patches update the same file in the same place, so whatever order you do them in, one chunk of the second patch you apply will fail. The botched solution is to re-apply the failed chunk of the second set of patches with totally lax context checking. This makes the patch in the wrong place, about 5 lines too early, but since both sets of patches are just adding #includes here it may just work.
The other kernel is my main one for nekon (and now nekon2 as well), including the openMosix patches on top of the existing squashfs ones to be able to see whether I really know how to get OM working or not. Applying the patches to my regular kernel sources fails (saying the patches have already been applied and squawking horribly) so it's back to another virgin 2.4.20 kernel, apply the patches and transport the .config from my main system kernel.
Compilations in progress...