[03:52] <flupke> hi, I'm looking for information on making my own linux/initrd.gz images for a netboot install
[03:53] <flupke> so far I have been able to boot the kernel, but how do I make the initrd start the installation script ?
[09:05] <cjwatson> flupke: I hope you aren't trying to do it with update-initramfs ...?
[09:06] <cjwatson> flupke: the netboot install initrds are built using the debian-installer package; if you use that then you would have to go to some effort to make it not start the installer on boot
[13:50] <flupke> cjwatson: heh I was using update-initramfs... thanks I did'nt know about debian-installer
[13:56] <xivulon> superm1 can you have someone to test mythbuntu rev 513+ in http://wubi-installer.org/devel/minefield?
[14:23] <abli_> How can I pass a preseed file to ubiquity if I am running ubiquity from the command line after booting the live CD? Also: running "ubiquity -d noniteractive" from a gnome terminal on the live cd crashes with "partam exited with code 10" being the last message. The disk has a large empty space on it. Any ideas what might be causing that?
[14:24] <cjwatson> abli_: pass the preseed file as input to 'sudo debconf-set-selections' first
[14:24] <cjwatson> abli_: running hardy?
[14:24] <cjwatson> abli_: sounds like bug 182004
[14:25] <abli_> ubuntu-8.04.1-desktop-amd64.iso in a kvm virtual machine. I'll check that bug
[14:47] <abli_> How is README.diskdefines file on the live cd used? what would happen if I would use the one from the i386 live cd on the amd64 live cd?
[14:50] <cjwatson> I don't think it's used by very much
[14:50] <cjwatson> I wouldn't worry about it
[14:51] <abli_> ok. thanks.
[15:28] <CIA-14> usb-creator: evand * r42 usb-creator/ (debian/changelog desktop/usb-creator.desktop): * Move from System Tools to Administration (LP: #285009).
[15:41] <CIA-14> usb-creator: evand * r43 usb-creator/debian/changelog: releasing version 0.1.9
[17:25] <CIA-14> base-installer: cjwatson * r335 ubuntu/ (3 files in 3 dirs): Add support for virtual kernel flavour on amd64 (LP: #283368).
[17:26] <superm1> evand, are you planning at least one more ubiquity upload before gold image?
[17:26] <cjwatson> there definitely needs to be
[17:27] <evand> indeed
[17:27] <superm1> okay i've got at least one critical mythbuntu thing to sneak into it then
[17:29] <CIA-14> base-installer: cjwatson * r336 ubuntu/debian/changelog: releasing version 1.86ubuntu7
[17:46] <CIA-14> ubiquity: superm1 * r2908 ubiquity/ (debian/changelog scripts/mythbuntu/mythbuntu_install.py): Add commit about broken cdrom entries to mythbuntu frontend
[17:46] <cjwatson> I have some more stuff on its way but may take a few hours
[17:46] <cjwatson> the langpack stuff still isn't working quite right :(
[17:46] <evand> noted
[17:47] <evand> I'm in grub-installer now, so I'm waiting on that as well.
[17:47] <CIA-14> ubiquity: superm1 * r2909 ubiquity/ (debian/changelog scripts/mythbuntu/apply-type): properly pre-setup database on mythbuntu frontend
[18:25] <kirkland> cjwatson: were you opposed to adding the _netdev bits to the partman fstab writer for iscsi?
[18:25] <kirkland> cjwatson: i'm trying to resolve the bits i saw in -meeting and your question in the bug
[18:30] <cjwatson> I don't mind if it's necessary
[19:05] <kirkland> cjwatson: i think it might be required, in conjunction with another change
[19:05] <kirkland> cjwatson: Etienne is testing a package I prepared for him
[19:05] <kirkland> cjwatson: in any case, Soren is supposed to be back at work on Monday, and I'd like his review of any iscsi changes I'd propose, as it's modifying work he previously started
[19:28] <james_w> persistence confirmed working on today's daily, thanks.
[19:28] <evand> hooray
[19:28] <james_w> I still get the errors on shutdown
[19:38] <james_w> end_request, I/O error, /dev/sdb, sector <some number>
[19:38] <james_w> repeated a lot with different sectors
[19:39] <james_w> between hitting enter when prompted to remove the CD and the machine shutting down
[19:40] <evand> hrm
[19:41] <evand> Does this also happen on the regular live CD?  I'm assuming sdb would be one of your regular disks.
[19:42] <james_w> Colin suggested it might be a problem with the order in which things are unmounted, but I can't see that casper has anything to do with unmounting, is it just the normal mechanism?
[19:42] <james_w> I'm just booting again to find out what /dev/sdb is
[19:42] <james_w> it doesn't seem to happen without persistence, I'll check again in a minute
[19:45] <james_w> /dev/sdb is the USB stick I am using
[19:46] <evand> odd, I would figure that would end up as sda.
[19:51] <james_w> yeah, no messages without persistence
[19:56] <james_w> I'm off to make dinner. If there is anything you would like me to try I will be happy to
[19:56] <evand> ok, I'll try to think through this one a bit more
[20:23] <evand> james_w: how are you reproducing this?  That is, aside from removing quiet splash from the kernel command line, is there anything else you're doing?
[20:23] <evand> -no-shutdown in kvm appears to be broken, unless I'm missing something
[20:44] <CIA-14> usb-creator: evand * r44 usb-creator/ (debian/changelog scripts/install.py):
[20:44] <CIA-14> usb-creator: Always write cdrom-detect/try-usb=true, rather than just when
[20:44] <CIA-14> usb-creator: persistence is enabled.
[20:48] <james_w> nothing else. ISO->USB with USB creator, with persistence. Add "persistence" to kernel command line, removing "quiet splash" so I can see what is going on, then booting and shutting down.
[20:48]  * james_w -> pub
[20:48] <evand> using KVM or real hardware?
[20:49] <james_w> real hardware. I'll try and find a second machine and USB to rule that out
[20:49] <evand> ah, I'll try with real hardware in a bit as KVM is failing me.
[21:28] <CIA-14> usb-creator: evand * r45 usb-creator/ (debian/changelog scripts/install.py):
[21:28] <CIA-14> usb-creator: persist was always getting enabled as the flag for it was being passed as
[21:28] <CIA-14> usb-creator: a string rather than an integer.
[22:29] <evand> hrm, cdrom-detect/ask-usb fails to account for the situation where you have a CD in the drive as well.
[22:31] <cjwatson> it's not clear what the user actually wants there :)
[22:31] <cjwatson> we'd have to detect BIOS boot order ...
[22:31] <cjwatson> (i.e. impossible)
[22:32] <evand> surely we could either A) try to unmount /cdrom before running the block inside db_get cdrom-detect/try-usb
[22:33] <evand> or B) (and scary for this late in the cycle) rework the code to check for a proper Ubuntu CD device by device.
[22:33] <evand> I suspect my original description of the problem may have been confusing
[22:34] <evand> Right now I have an Ubuntu image on USB stick plugged in and a Windows drivers CD inserted into the computer.
[22:34] <evand> It first mounts the CD, then goes into the try-usb block and mounts the usb disk over top of that.
[22:35] <evand> If I stick a if [ "$mounted" = "1" ]; then break; fi before try-usb it goes with the Windows CD, which obviously fails.
[22:36] <cjwatson> oh, hmm
[22:36] <evand> I suspect in this scenario, it should try to mount, then check to make sure we're dealing with an Ubuntu image, if not go to the next device, if there are no devices left, fail
[22:37] <cjwatson> I see what you mean, that's badly wrong
[22:38] <cjwatson> I'm not sure as to the right answer, though, buried in fixing language pack stuff right now
[22:38] <cjwatson> fix it as you see fit :)
[22:38] <evand> indeed, I'll see what I can come up with
[22:41] <cjwatson> think I might have nailed langpacks now
[22:44] <cjwatson> several problems: (1) wasn't deciding which language packs to use before generating the blacklist (2) was filtering out packages with prerms before doing ok-to-remove calculations (3) needed to try recursively removing broken packages because apt is a bit weird (4) was removing packages from the difference list too early so they stopped being considered as packages that were allowed to be removed (er, I think)
[22:48] <evand> ahh
[22:48] <evand> wow, thanks