[00:27] I'm building a custom fully-automated, but I can't find what's the d-i attribute to avoid it from using the network to install packages... is there an option to make it use the cdrom only, or maybe a hack to set it as a mirror? [00:28] I have "d-i apt-setup/use_mirror boolean false" set, but still not working [00:54] cjwatson: do you know if that kludge is still worth using or is this issue fixed? http://irclogs.ubuntu.com/2006/12/08/%23ubuntu-installer.txt [01:26] cjwatson: I'm having trouble with hardy alpha-3's casper writing resolv.conf. In 23networking, /domain/ matches no lines in /netboot.config [01:26] (I forget, are you a casper hacker?) [01:29] cjwatson: well... it works... but still feel a bit hackish [01:32] twb: Have you tried alpha 5? [01:32] It wasn't around last time I looked. [01:32] Alpha 5 was released last week. [01:33] I might actually be using a4 now, lemme check [02:56] sooo, anyone awake in here? [03:15] Yes I am, but I am not one of the primary maintainers of the installer. [03:15] ah. Well, would you be up for a question? [03:16] Ok, I'll see if I can help. [03:17] It's a pretty odd one. Basically, my CD boots properly, but as soon as the kernel begins to boot from the CD (the usplash with the bar that slides back and forth) it starts toggling my floppy drive. [03:17] It tries to access 19 times, and then dumps me to a BusyBox prompt. [03:19] If I put a floppy in the drive, it tries to access the floppy for about 10 seconds, then proceeds to boot normally. If I disable the floppy drive in the BIOS, I run into the same BusyBox prompt. [03:22] Chorca: Well thats easy so far as working out why the floppy drive is being accessed. FOr some reason, the initial boot stages of the CD cannot find your CD-ROM drive. What version of Ubuntu is this by the way? [03:23] Hardy Alpha 5 [03:23] this also happened in Alpha 4 [03:23] Ah ok. Does Ubuntu gutsy boot ok? [03:24] I haven't tried.. I think i have a CD around here, let me get it real quick. [03:24] Ok. Did you try hardy alpha 3 by any chance? [03:25] Nah, i started in on alpha 4 [03:25] Hm ok. [03:25] I would try Ubuntu gutsy first, so we know whether there is a regression. [03:25] I've got a gutsy CD here, booting it now. [03:28] Gutsy also toggles the drive, though slower and about half as many times, but then proceeds to boot normally from the CD. [03:28] Ok, sounds like your hardware is not being recognised properly. Have you run any Ubuntu on this system at all int he past? [03:28] in the past even [03:29] Well I used to run Dapper I think, it's been awhile. It's an old system I put back into service recently. [03:29] Gigabyte GA-7NNXP board and AMD Athlon 2800+ processor [03:31] heh, though it appears gutsy doesn't like the nvidia card and DVI monitor, there's no video out once it's booted. [03:32] oh sorry, didn't read what you said... Right then we do have a regression somewhere. [03:33] Since I don't know how the CD boots that well, I suggest you file a bug in launchpad against the casper package, describing what you ahve tried, and I'll see about getting someone to look at it who knows more about how the bits work. [03:33] only other things on it are a Silicon Image SATA 150 controller, and an ITE IDE RID. [03:33] RAID* [03:33] okay. I started a thread in the Hardy forum as well on it. [03:33] Right, well the best place to get developers to see it is filing a bug. [03:34] alright, i'll start a bug up onnit. [03:34] Chorca: Thanks. [03:39] ok, call me silly, I can't seem to find a place in casper for bugs. Or should I just put a general bug in. [03:39] Chorca: Go to https://launchpad.net/ubuntu/+source/casper/+bugs click report a bug, and you will find the package field already filled in for you. [03:40] ooh there we go [03:45] https://bugs.launchpad.net/ubuntu/+source/casper/+bug/195993 fiiiiled [03:45] Launchpad bug 195993 in casper "Install CD dumps to busybox after toggling floppy" [Undecided,New] [03:46] ell oh ell [03:46] Chorca: Thanks. [03:53] Thanksthanks for the help, too. [03:53] :) [04:08] Chorca: Welcome. === Chorca1 is now known as Chorca [09:37] hi [09:37] cjwatson: ping ? [09:38] i move the debian installer on a a new machine which is connected to a SAN and i have somme mounts on this and i get some error such as an iso still building but the lockfile exists, etc... have you an idea ? [11:16] cjwatson: You were right, dmraid45 did make it into the kernel, I just wans't looking in the usual module locations. :) [13:32] Have a few questions: [13:32] 1) where do I find preseeding values for supported keyboard variant codes? [13:33] 2) What do you think about 192116? [13:33] /usr/share/console-setup/KeyboardNames.pl, right-hand side of each element of the %variants hash [13:33] they originate in xkb-data [13:34] 3) Edubuntu devs suggested to support the addon ISO in wubi, what would be a good way to that? apt-cdrom was mentioned [13:34] 2) I don't see why users shouldn't just be re-educated. Mounting in /media/host is IMO wrong and likely to create problems. [13:34] 3b) can I use apt-cdrom with an ISO? at what stage shall I call it? early command and success command seem my only options [13:35] 3) the only thing I think you can do with umenu on the Edubuntu add-on CD is to recommend that they install Ubuntu first [13:35] isn't apt-cdrom a viable route then? [13:36] no, you still have to install Ubuntu first [13:36] no way round that [13:36] yeah the idea would be to download both the Ubuntu ISO + addon ISO [13:36] boot off the Ubuntu ISO, adding the addon ISO as repo [13:36] then installing edubuntu-desktop [13:37] err. I don't really know [13:37] "adding the addon ISO as repo" is the part I am not sure how to do (from within preseeding) [13:37] I'd be surprised if there was all that much call for it in practice [13:37] I would agree with that [13:37] you can give apt-cdrom an arbitrary mount point with -d though, and loop-mount something there first [13:38] does it have to be on fstab? [13:38] I don't think so provided that you also use -m) [13:38] -m [13:39] the next issue then is how to call apt-cdrom from within a preseed file [13:39] I assume early command is too soon and success command is doable but not ideal [13:39] don't think I have other hooks available [13:40] early_command is too soon, yes [13:41] ubiquity: cjwatson * r2506 ubiquity/ (debian/changelog ubiquity/filteredcommand.py): [13:41] ubiquity: * Print a traceback as well as the exception message if [13:41] ubiquity: DebconfFilter.process_line raises an exception. [13:41] TBH I don't think it's worth the necessary contortions. People can just install Ubuntu and then insert the Edubuntu add-on CD later on. [13:42] we aren't going to support direct installation of the add-on CD via any other method [13:43] I liked the idea, as suggested by stgraber, also because it would be an avenue to add support to other distros now only available via the alternate CD [13:44] but that of course can be postponed to hardy+1 [13:46] I'd assume that is your preferred option, hence I'll remove edubuntu from wubi [13:46] I think so, yes, but run it by the Edubuntu guys with rationale [13:46] will do a pastebin of the above [13:48] ubiquity: cjwatson * r2507 ubiquity/ (6 files in 3 dirs): [13:48] ubiquity: * Enable debconf's escape capability for communications between the filter [13:48] ubiquity: and the underlying frontend by default. Split arguments to filtered [13:48] ubiquity: commands into only as many arguments as are required by the command. [13:48] ubiquity: This should fix autopartitioning in certain locales (LP: #177623). [13:48] evand: another pair of eyes on r2507 wouldn't hurt. If you encounter weird problems with debconf filtering then it's probably down to that, one way or another. However, it seems to be working better for me so far ... [13:53] evand: the main symptom was that double spaces weren't always surviving filtering, because they got eaten by argument splitting; it turned out that the most straightforward way to fix that involved making filter<->frontend communication always use the escape capability, but that involved some care to make sure that we didn't blat escape on for confmodules that aren't expecting it [14:11] apt-setup: cjwatson * r125 apt-setup/ (debian/changelog finish-install.apt-cdrom-setup): [14:11] apt-setup: * Don't bother running 'apt-get update' after disabling cdrom entries in [14:11] apt-setup: sources.list; it can be slow and is unnecessary. [14:13] apt-setup: cjwatson * r126 apt-setup/debian/changelog: releasing version 1:0.31ubuntu6 [14:19] cjwatson thanks for the answers ^ [14:22] cjwatson: ok, I'll take a look [14:37] cjwatson: do you think splitting failure_command into failure_command and automation_failure_command (user input needed) falls under UI freeze? [14:38] hrm, actually nevermind. That's not that big a change and should only take me a few minutes today. [14:39] re edubuntu umenu, my understanding is that there is no stand-alone edubuntu ISO so there will be no "edubuntu umenu" either [14:40] so no UI changes there (we can leave the artwork in anyway) [15:15] evand: no, I don't think so [15:16] ok [17:04] ubiquity: cjwatson * r2508 ubiquity/ubiquity/debconffilter.py: don't re-escape command arguments sent by confmodules with the escape capability set [17:04] cjwatson: 2507 obviously looks reasonable, but I'll keep an eye out for side effects in new bug reports. [17:04] ah, I spoke too soon apparently :) [17:05] heh, just broke the summary page a bit [17:24] ubiquity: evand * r2509 ubiquity/ (6 files in 2 dirs): [17:24] ubiquity: * Add ubiquity/automation_failure_command debconf question which is called [17:24] ubiquity: whenever user interaction is required. This includes non-fatal error [17:24] ubiquity: dialogs. ubiquity/failure_command is now only called when the installer [17:24] ubiquity: crashes. [17:36] cjwatson: why shebang in shell script are /bin/sh ? it some use some bash functions such as pushd, etc... [17:40] which particular shell scripts? [17:40] if a script uses bashisms but is #! /bin/sh then that is unambiguously a bug and should be filed as such [17:41] but I know of no installer shell scripts that do that [18:05] bdmurray: weirdly, I can't find the fix for 89528 (or rather I can, but it predates the version the reporter is using), so I don't feel entirely comfortable closing it. Instead I've just unassigned myself from it, so if he still doesn't reply then the Launchpad Janitor will close it eventually. [18:14] superm1: have you seen bug 196199 before? looks slightly different from the previous similar ones; I don't know if it's the same as 149935 and 150029 [18:14] Launchpad bug 196199 in ubiquity "Issue during Mythbuntu install from live cd 07.10" [Undecided,New] https://launchpad.net/bugs/196199 [18:18] cjwatson: okay, thanks. Additionally, I was doing some testing of the minimum memory requirements for Hardy and put notes in bug 105682. What more should we do? [18:18] Launchpad bug 105682 in casper "[gutsy] Tribe 5 amd64 needs more than 256MB of memory" [Medium,Confirmed] https://launchpad.net/bugs/105682 [18:33] bdmurray: (a) hardy-targetted task on debian-installer to adjust help text (b) would also be useful to have figures for the boot mode that goes straight into the installer [18:33] possibly even more useful [18:34] 512MB is huge as minimum requirements go. I'd be a lot happier with 384MB as a minimum requirement, and happier still if 256MB could be enough for install-only mode. [18:36] right, I got 256MB to work with only-ubiquity yesterday w/ swap and evand said it didn't work w/o swap for only-ubiquity [18:36] i.e. creating a swap partition during installation? [18:38] 13:47 < evand> bdmurray: ugh. ubiquity fails in 256M under those conditions as well (no swap before or during install). [18:38] I think it's fine to require one to be created during installation [18:53] ah, if that's reasonable then 256MB is adequate for an install in only-ubiquity [19:10] xivulon: wubi and umenu do not work with the kde4 kubuntu CDs: http://cdimage.ubuntu.com/kubuntu-kde4/ [20:02] evand: it's similar to the requirements that d-i has traditionally used, at least [20:04] cjwatson: fair enough, I just didn't want someone to come back and say, "it didn't work in this configuration despite what the requirements told me", but if you're confident in it then I am as well. [20:32] ubiquity: cjwatson * r2510 ubiquity/ubiquity/components/mythbuntu_apply.py: bug number typo [20:47] evand what is the issue exactly? [20:47] we also have another problem: http://ubuntuforums.org/showpost.php?p=4416099&postcount=85 [20:48] basically sometime device.map and grub4dos do not agree on the disk ordering [20:48] I'll need some help with that! [20:49] we have this problem in Linux as well. As I understand it, there's no easy solution. [20:50] ah ok then I am happier :P [20:50] regarding the Kubuntu KDE4 issue, it's a matter of the .disk/info string being different, I believe [20:51] bugger [20:51] Kubuntu-KDE4 8.04 "Hardy Heron" - Alpha amd64 (20080227.2) [20:51] was about to ask that [20:51] do you have a Ubuntu one handy too? [20:51] ah I know [20:51] the - in the name! [20:52] I use "-" as a delimiter together with whitespace [20:52] ok [20:52] will fix that [20:55] bug 196262 [20:55] Launchpad bug 196262 in wubi "Wrong parsing of kubuntu .disk/info" [Medium,Confirmed] https://launchpad.net/bugs/196262 [20:58] re device.map, I hope that the upstream fix will involve device.map side of things as opposed to fix grub booting code [20:58] otherwise we will have to port the changes to grub4dos [21:03] is there a bug for that by the way? Did a quick search but I cannot find it. [21:13] bug 8497 [21:13] argh, he left [21:13] Launchpad bug 8497 in grub "grub guessed BIOS disk order incorrectly" [Medium,Confirmed] https://launchpad.net/bugs/8497 === cr3_ is now known as cr3