[01:00] <DivGradCurl> Anyone here?
[01:02] <lifeless> this is really a development channel
[01:03] <lifeless> I suspect you are better asking on #ubuntu-users
[01:04] <DivGradCurl> Maybe, maybe not.  I'm not 100% clear what constitutes a development question/issue, but there is a decent chance that my question *is* development related.
[01:04] <lifeless> well, ask the question rather than asking to ask ;)
[01:04] <lifeless> then folk can advise
[01:05] <DivGradCurl> There has emerged advice on how to load Ubuntu onto external USB drives (http://ubuntuforums.org/showthread.php?t=80811), but I haven't been able to dig up how to Boot Ubuntu from an external USB drive using a computer whose BIOS doesn't support USB booting.
[01:05] <DivGradCurl> Other distros have instructions and the capability, but there doesn't seem a clearcut method for Ubuntu.
[01:06] <DivGradCurl> A year-old article for Puppy exists: http://www.goosee.com/puppy/boot2pup.htm
[01:06] <lifeless> you mean something like 'put in a floppy with the kernel' ?
[01:07] <DivGradCurl> Like I said, I'm new, and have no idea how to do something like that.  And this may be the wrong place to ask how to do so.
[01:08] <DivGradCurl> Just figured it was Boot-related and has general application since Ubuntu isn't as geared towards setting this (or another method) up as have other distros.
[01:09] <DivGradCurl> My general impression is that a major problem is that it must first DETECT the USB capability before it can consider trying to boot from it.
[01:10] <DivGradCurl> I do have Ubuntu installed on a local computer (~3.5 years) that doesn't have USB Booting.  I tried editing its grub to boot from hd1 or sda but haven't had any luck.
[01:11] <DivGradCurl> Then again, it takes 3 seconds for the computer to hit the grub screen so I have no idea if (a) I am doing anything comprable to having a kernel on a disc and (b) if so, whether I am doing it wrong or if having a kernel & grub is insufficient.
[01:18] <lifeless> the problem is that to boot from USB, you -must- have the computer load the USB disks bootsector
[01:18] <lifeless> this is not at all the same as having / on USB on a computer that cannot boot from USB.
[03:31] <jbailey> The new klibc upload moves klibc to /lib, but retains the old symlink.
[03:31] <jbailey> Sideffect: you can run klibc binaries on the running system.
[03:39] <fabbione> that's good right?
[03:40] <jbailey> Yes, it means that using klibc in an application sucks a bit less.
[03:44] <Keybuk> is anything using klibc now? :)
[03:46] <jbailey> No, but if you chose to, it would now suck less.
[03:47] <Keybuk> except for the bludgeoning pain of compiling anything against it, and having to do all that copying and pasting of linux headers
[03:47] <jbailey> Compiling against it?
[03:47] <jbailey> Just use klcc. =)
[03:48] <jbailey> But yeah, the linux headers are next on the list to fix.
[03:48] <jbailey> Next week, maybe.
[03:48] <jbailey> It'd be before I go to the UK, though.
[03:48] <Keybuk> did you get that glibc nameserver patch in yet?
[03:48] <jbailey> Keybuk: No.  Next upload should have it.  I wanted 2.3.6 to just go in.
[03:48] <Keybuk> ok
[03:48] <jbailey> It's a big enough set of changes on its own.
[03:49] <Keybuk> n-m is looking "likely" again
[03:49] <Keybuk> so it'd be nice to have
[03:49] <jbailey> I noticed that.
[03:49] <jbailey> I was going to comment on behalf of one of your use cases, thank-you.
[03:49] <jbailey> =)
[03:49] <Keybuk> heh
[03:50] <Keybuk> it'd just mean that in Berlin? we'd have more specs, along the lines of "support static IP configuration per-network", "support PPP", "support VPN", etc.
[03:50] <jbailey> Right.
[03:50] <jbailey> Since clearly "do network manager" would be the wrong thing.
[03:52] <Keybuk> well, is more that n-m can support those things, but I'm not going to bother testing it for dapper or trying hard not to break those things
[03:52] <Keybuk> in particular, static IP won't work
[09:56] <fabbione> Keybuk: for the sake of saving me some pain
[09:56] <fabbione> where do i need to install udev rules?
[09:56] <fabbione> http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/dlm/lib/51-dlm.rules.diff?cvsroot=cluster&only_with_tag=STABLE&r1=NONE&r2=1.1.2.1
[09:56] <fabbione> they just added the rules for cluster stuff
[09:57] <Keybuk> /etc/udev/rules.d
[09:58] <fabbione> Keybuk: do they look sane?
[09:58] <Keybuk> there's a bug in those rules
[09:58] <fabbione> such as
[09:58] <Keybuk> needs to be NAME== not NAME=, and needs to be a comma before MODE=
[09:58] <fabbione> ok thanks
[09:58] <Keybuk> +KERNEL=="dlm-control", NAME=="misc/dlm-control"
[09:58] <Keybuk> +KERNEL=="dlm_default", NAME=="misc/dlm_default", MODE="0666"
[09:58] <Keybuk> +KERNEL=="dlm_*", NAME=="misc/%k", MODE="0660"
[09:58] <Keybuk> like that
[09:58] <Keybuk> install them as
[09:58] <fabbione> yeps
[09:59] <Keybuk> /etc/udev/rules.d/45-dlm.rules
[09:59] <Keybuk> though, technically, if you want to be absolutely correct with Ubuntu scheming
[09:59] <Keybuk> you should put the KERNEL==/NAME== bits in /etc/udev/rules.d/25-dlm.rules
[09:59] <Keybuk> and the MODE= bits in /etc/udev/rules.d/45-dlm.rules
[09:59] <fabbione> meh
[10:00] <Keybuk> though if there's no chance of this being needed in initramfs, or in the installer, then don't worry too muhc
[10:00] <fabbione> no, they don't need to
[10:00] <Keybuk> ok
[10:00] <fabbione> thanks a lot
[10:01] <Keybuk> hmm
[10:01] <Keybuk> those last two rules look bogus ...
[10:01] <Keybuk> +KERNEL=="dlm_default", NAME="misc/dlm_default" MODE="0666"
[10:01] <Keybuk> +KERNEL=="dlm_*", NAME="misc/%k" MODE="0660"
[10:01] <Keybuk> dlm_default matches dlm_*, so it'll get 0660, not 666
[10:01] <Keybuk> swap those two around
[10:02] <fabbione> in theory one is enough...
[10:02] <fabbione> isn't it?
[10:02] <fabbione> just the last line
[10:02] <Keybuk> in theory you just need:
[10:02] <Keybuk> KERNEL=="dlm_*", NAME="misc/%k"
[10:02] <Keybuk> KERNEL=="dlm_default", MODE="0666"
[10:03] <fabbione> why the last one?
[10:03] <Keybuk> to change the mode from 0660 to 0666
[10:03] <fabbione> oh ok
[10:04] <fabbione> pjc thanks very much.
[10:04] <fabbione> pjc oddly they still seem to work :)
[10:04] <fabbione> fabbione oh halt.. we can make it very very simple
[10:04] <fabbione> fabbione Keybuk KERNEL=="dlm_*", NAME="misc/%k"
[10:04] <fabbione> fabbione Keybuk KERNEL=="dlm_default", MODE="0666"
[10:04] <fabbione> fabbione this 2 lines are the "correct" equivalent of the 3 you have now :)
[10:04] <fabbione> pjc no, cos dlm-control has a hypen, not an underscore
[10:04] <fabbione> fabbione oh right
[10:04] <fabbione> pjc (so i doesn't get confused with lockspace names)
[10:05] <fabbione> fabbione right
[10:05] <fabbione> fabbione i missed that
[10:05] <fabbione> fabbione thanks
[10:05] <fabbione> Keybuk: i already pushed the fix upstream :)
[10:05] <fabbione> but he is right.. you need 3 rules
[10:05] <Keybuk> yeah, sorry, you need the first one too
[10:05] <fabbione> yup
[10:05] <Keybuk> depends which udev version you have, my rules will work with old and new udev, his only with old
[10:05] <fabbione> so he sais
[10:14] <fabbione> KERNEL=="dlm-control", NAME=="misc/dlm-control"
[10:14] <fabbione> KERNEL=="dlm_default", NAME=="misc/dlm_default", MODE="0666"
[10:14] <fabbione> KERNEL=="dlm_*", NAME=="misc/%k", MODE="0660"
[10:14] <fabbione> so like this
[10:14] <fabbione> or do i still need to swap the last 2?
[10:15] <Keybuk> still need to swap the last two, otherwise the match for the last one will change the permissions
[10:15] <fabbione> ok
[10:15] <Keybuk> dlm_* matches dlm_default
[10:15] <Keybuk> KERNEL=="dlm-control", NAME=="misc/dlm-control"
[10:15] <Keybuk> KERNEL=="dlm_*", NAME=="misc/%k", MODE="0660"
[10:15] <Keybuk> KERNEL=="dlm_default", MODE="0666"
[10:15] <Keybuk> is how I'd do it
[10:16] <Keybuk> (dlm_default is renamed by the middle one)
[10:16] <fabbione> yup
[10:16] <fabbione> thanks