**** BEGIN LOGGING AT Thu Mar 30 10:59:56 2006
Mar 30 13:15:33 <Snider> NAiL, I didn't use shutdown, if the slug loses power, it will not boot back up
Mar 30 14:05:17 <caplink811_log> Snider, maybe, that http://www.nslu2-linux.org/wiki/FAQ/DiskNotMountedAtBoot help you
Mar 30 15:11:36 <NAiL> I've got problems getting TCL to compile: http://pastebin.com/630886
Mar 30 15:12:11 <NAiL> I think this might be a ubuntu-specific problem, I just can't figure out what
Mar 30 15:16:33 <mwester> I didn't know you ran ubuntu...
Mar 30 15:16:49 <NAiL> I do ;)
Mar 30 15:17:13 <NAiL> but the configure source looks ok. Looks like a perfectly ok case-statement
Mar 30 15:17:22 <mwester> That's a standard shell script AFAIK...
Mar 30 15:18:12 <NAiL> I'm gonna get my box up-to-date
Mar 30 15:29:25 <mwester> You could try Fedora Core 5 -- that's up to date!   :)
Mar 30 17:16:03 <likewise> hi all
Mar 30 19:12:12 <Snider> caplink811_log, thanks, did not see that.
Mar 30 19:15:31 <caplink811_log> np
Mar 30 19:15:36 <Snider> But I'm running debian slug not unslung/openslug... I will try to remove the .recovery file if it does it again
Mar 30 19:16:01 <Snider> I successfully issued the reboot command, but did not try shutdown -h now yet
Mar 30 19:16:37 <Snider> If it doesn't boot after shutdown -h now there is a problem.
Mar 30 19:16:37 <caplink811_log> imho, the mechanism is the same
Mar 30 19:16:59 <Snider> I will have to mount that drive to my PC to remove the .recovery file
Mar 30 19:17:57 <EvilDevil> Snider you can boot without the disk, ssh into the nslu2, plug in the disk,mount it and remove the .recovery
Mar 30 19:41:11 <likewise> all: who of you build from OE?
Mar 30 19:41:39 <NAiL> I built from OE today
Mar 30 19:41:52 <NAiL> I didn't get all the packages built though, since tcl fails on my box
Mar 30 19:41:54 <EvilDevil> OE + CVS
Mar 30 19:43:35 <likewise> NAiL: could you please tell me what your local.conf looks like, I ran into 5 different errors the last few days, from GCC ICE bugs to missing dependencies, etc.
Mar 30 19:44:23 <NAiL> actually, I don't have a local.conf. I'm building the stock debianslug image
Mar 30 19:45:47 <likewise> NAiL: How do you tell bitbake to run against debianslug.conf?
Mar 30 19:46:21 <NAiL> repvik@Ximinez:~/slug/debianslug/conf$ cat auto.conf
Mar 30 19:46:21 <NAiL> DISTRO="debianslug"
Mar 30 19:46:21 <NAiL> MACHINE="nslu2"
Mar 30 19:46:47 <NAiL> That's the only thing that differs between the different slugos variants config-wise
Mar 30 19:47:03 <NAiL> (not in openembedded/conf/distro, of course)
Mar 30 19:47:17 <NAiL> morning, rwhitby 
Mar 30 19:47:29 <rwhitby> morning
Mar 30 19:47:30 <likewise> Hi rwhitby
Mar 30 19:47:37 <NAiL> rwhitby: I've proposed to change the way slugos boots, to make the process easier to manage
Mar 30 19:48:00 <rwhitby> is it in the logs?
Mar 30 19:48:04 <NAiL> yes
Mar 30 19:48:09 <likewise> NAiL: Uhm, so I am missing something really big here, since I *am* using org.openembedded.dev/conf/distro files
Mar 30 19:49:05 <NAiL> rwhitby: #openslug, if you're wondering
Mar 30 19:50:42 <NAiL> rwhitby: It's just too messy the way it is today. Any changes to /linuxrc are overwritten if you turnup again.
Mar 30 19:53:08 <rwhitby> Well, people aren't usually meant to be editing /linuxrc directly - that's for power users only.
Mar 30 19:53:38 <rwhitby> what are you needing to edit in there as an end-user (as opposed to a slugos developer getting it to work with modularised usb for the first time) ?
Mar 30 19:54:43 <NAiL> linuxrc currently beeps, sets leds, loads modules and then boots
Mar 30 19:54:56 <NAiL> managing linuxrc is a PITA, since it's echoed from turnup
Mar 30 19:55:21 <NAiL> IMHO, the only part that should be echoed from turnup is the mount command itself
Mar 30 19:55:39 <NAiL> (ie, "/boot/disk <UUID> -t etc")
Mar 30 19:56:44 <NAiL> that way, only the /boot/initial (or whatever one would like to call it) would be changed. And that would only be one line.
Mar 30 20:00:08 <NAiL> For me as a dev, it makes it easier to fix problems while it allows my custom changes as a user to stick
Mar 30 20:00:19 <NAiL> s/user/power user/
Mar 30 20:00:19 <jbot> NAiL meant: For me as a dev, it makes it easier to fix problems while it allows my custom changes as a power user to stick
Mar 30 20:09:47 <rwhitby> NAiL: ok, a /boot/initial which is included from /linuxrc is a good idea.
Mar 30 20:10:32 <NAiL> What I haven't quite figured out yet is how it will be when users upgrade
Mar 30 20:11:23 <rwhitby> The /linuxrc will stay as it is until the user does a turnup
Mar 30 20:11:33 <rwhitby> I don't think upgrades touch /linuxrc at all.
Mar 30 20:12:13 <rwhitby> the upgrade would put the /boot/initial in place, and would update /boot/disk, so people would loose the stuff which was moved to /boot/initial until they did a new turnup disk
Mar 30 20:12:32 <NAiL> hmm
Mar 30 20:12:43 <NAiL> yes
Mar 30 20:12:45 <NAiL> makes sense
Mar 30 20:12:56 <rwhitby> So perhaps include /boot/initial from /boot/disk instead.
Mar 30 20:13:25 <NAiL> uh, that wouldn't make much sense
Mar 30 20:13:30 <rwhitby> then the user's /linuxrc doesn't need to change, and the worst that can happen is that modprobe goes twice.
Mar 30 20:13:49 <rwhitby> (and the leds get set twice)
Mar 30 20:13:54 <rwhitby> what doesn't make sense?
Mar 30 20:13:57 <NAiL> boot/initial decides (written by turnup) which medium (disk, flash, ram, nfs) one boots to
Mar 30 20:14:09 <rwhitby> whoa, hang on there.
Mar 30 20:14:17 <rwhitby> turnup writes /linuxrc, nothing else.
Mar 30 20:14:43 <NAiL> hmm
Mar 30 20:14:44 <NAiL> ok
Mar 30 20:14:45 <rwhitby> turnup must *not* write to any file under /boot
Mar 30 20:15:11 <NAiL> fair enough, linuxrc is changed to "exec /boot/initial ; exec /boot/whateveritwasbefore"
Mar 30 20:15:23 <rwhitby> right.
Mar 30 20:15:45 <rwhitby> or you get /boot/disk to exec /boot/initial
Mar 30 20:16:00 <rwhitby> (and /boot/flash etc to also exec it)
Mar 30 20:16:03 <EvilDevil> hmm /slug/openembedded/packages/slugos-init/slugos-init_0.10.bb says something about /linuxrc
Mar 30 20:16:30 <rwhitby> BTW, /linuxrc is usually just a symlink to a file in /boot
Mar 30 20:16:57 <rwhitby> no, that's wrong - that's only for the flash case
Mar 30 20:17:06 <NAiL> yes
Mar 30 20:17:55 <rwhitby> So if you want to split out initial stuff that is common to all /boot/<variant> files, then put that stuff in /boot/initial and exec it from each of the boot <variant> files.
Mar 30 20:18:07 <rwhitby> then your refactoring is invisible to the end user.
Mar 30 20:18:15 <rwhitby> and there are no upgrade problems
Mar 30 20:18:20 <NAiL> yes
Mar 30 20:18:40 <NAiL> and turnup only needs a tiny modification (removing crud)
Mar 30 20:18:44 <rwhitby> right
Mar 30 20:19:19 <rwhitby> as long as that crud is safe to execute twice (for the period where someone has upgraded, but not done a new turnup), then there are no upgrade problems
Mar 30 20:19:48 <rwhitby> setting leds, beeping, and modprobes should all be safe for multiple execution
Mar 30 20:19:55 <NAiL> yeah
Mar 30 20:19:59 <rwhitby> and the double beep will remind people to do a turnup to fix it.
Mar 30 20:20:08 <NAiL> there's a depmod in there too, but only if modules.dep doesn't already exist
Mar 30 20:20:56 <NAiL> ok, I'll move my changes to "initial", finish it up and commit
Mar 30 20:21:43 <NAiL> I'm considering undoing the modularisation in 3.6-beta though
Mar 30 20:22:10 <NAiL> I think the changes will be too intrusive (and not well enough tested) for a release
Mar 30 20:23:25 <rwhitby> nod
Mar 30 20:25:13 <NAiL> There's one issue there though... usb and scsi will have to go back in... How about libata-pata? Not entirely sure there's enough space for all of them
Mar 30 20:28:17 <EvilDevil> how many alpha testes would be needed to get the modularised kernel to *stable* beta?
Mar 30 20:28:29 <EvilDevil> s/testes/testers/
Mar 30 20:28:29 <jbot> EvilDevil meant: how many alpha testers would be needed to get the modularised kernel to *stable* beta?
Mar 30 20:29:37 <NAiL> dunno, but the way it looks now it's gonna take a while before it gets to the point where there's any use in testing it
Mar 30 20:30:06 <NAiL> I've got way too much to do, and my exam is coming up fast
Mar 30 20:33:35 <EvilDevil> exams go first
Mar 30 20:39:25 <eFfeM> nail if you are talking about modules, i have all  the usb drivers as y in my defconfig and that one still fits
Mar 31 00:32:02 <Snider> Ok, debian is reinstalled
Mar 31 00:32:18 <Snider> going to try rebooting / shutdown.
Mar 31 00:32:31 <Snider> I know reboot works fine, not sure about shutdown tho
Mar 31 01:07:04 <Daluv> NAiL: alive?
Mar 31 01:13:50 <NAiL> uh
Mar 31 01:13:51 <NAiL> yes
Mar 31 01:13:52 <NAiL> well
Mar 31 02:06:21 <Snider> NAiL, you experienced with debianslug?
Mar 31 02:06:33 <NAiL> hmm, a bit
Mar 31 02:07:31 <Snider> Well, I don't believe I can boot up with out sda1 with debian slug
Mar 31 02:09:56 <Snider> going to unplug it again
Mar 31 02:10:04 <Snider> and try to delete the .recovery file...
Mar 31 02:13:11 <NAiL> bit hard to say where it stops... have you tried just letting it be for a while? 
Mar 31 02:13:24 <Snider> over night
Mar 31 02:13:27 <NAiL> My slug once took a bit more than one hour ;)
Mar 31 02:13:34 <NAiL> oh... then it should've booted
Mar 31 02:13:46 <Snider> Yea
Mar 31 02:14:28 <NAiL> I assume you don't have serial?
Mar 31 02:15:09 <Snider> Working on it
Mar 31 02:16:40 <Snider> Will leave it overnight just to make sure it doesn't boot
Mar 31 02:17:17 <Snider> and try removing the .recovery file
Mar 31 02:17:23 <NAiL> big disk?
Mar 31 02:17:28 <Snider> 160GB
Mar 31 02:17:51 <NAiL> does it appear to work when you're booting?
Mar 31 02:17:56 <NAiL> eg. fsck?
Mar 31 02:19:39 <Snider> no
Mar 31 02:19:51 <Snider> At first it does, but then all HD activity stops
Mar 31 03:52:52 <eno-away> cvs.sf.net down, right?
Mar 31 05:05:51 <Snider> looks like its not coming back up
Mar 31 05:08:20 <Snider> and there is no .recovery in the root.
Mar 31 05:10:18 <Snider> So, what to do now.... seems like that is a problem :(
Mar 31 05:13:32 <Snider> last line in /var/log/dmesg is:
Mar 31 05:13:36 <Snider> Adding 88316k swap on /dev/sda5.  Priority:-1 extents:1 across:88316k
Mar 31 05:13:36 <Snider> EXT3 FS on sda1, internal journal
Mar 31 05:34:09 <Snider> This scares me: Jan  1 01:01:39 LKG7E3446 kernel: No module symbols loaded - kernel modules not
Mar 31 05:34:09 <Snider> enabled.
**** ENDING LOGGING AT Fri Mar 31 10:59:57 2006