[00:47] cjwatson, evand I think we'll have to skip dm-loop for the time being: http://img205.imageshack.us/my.php?image=dmloop3pj5.png [00:47] I will have a quick word with cking tomorrow just in case [00:52] yeah, that's similar to what I saw I think [01:06] strangely enough in the local code (which should be identical to the interactive commands that generated the kernel panic) I can go past the dmlosetup stage (although I run in the other problems mentioned above) [01:06] local code = /usr/share/initramfs-tools/scripts/local [01:15] cjwatson: I'm trying to debug the failure to install grub on a dmraid array from netboot, and have found a couple of things. 1) the error only occurs when installing alongside windows. 2) installing via netboot onto standard disks, i.e no dmraid array alongside windows causes grub installation to hang. [01:15] cjwatson: I'll ahve logs/more data to examine later today when you are around. [02:45] Hello. I asked a question on the forum and the Ubuntu Users list about preseeding dpkg-reconfigure console-setup but I didn't get any responses so I suspect it was the wrong place to ask. [02:46] Here is the forum post: http://ubuntuforums.org/showthread.php?t=943668 [02:46] Any help would be appreciated. [04:12] hey guys, I'm having some trouble with debian-cd [04:12] the error is: mv: cannot stat `/home/acoc/projects/ubuntu_devel/cdimage/work/scratch/ubuntu/daily-live/tmp/hardy-i386/CD1/casper/filesystem.kernel-generic': No such file or directory [04:12] make: *** [/home/acoc/projects/ubuntu_devel/cdimage/work/scratch/ubuntu/daily-live/tmp/hardy-i386/bootable-stamp] Error 1 [04:13] does anyone know what makes filesystem.kernel-generic or the initrd [04:19] acoc: livecd-rootfs I would think. [04:19] But I don't know for sure. [04:20] Yes, livecd-rootfs [04:21] It's the kernel from whatever linux-generic installs [04:25] sorry for the delay, I'm a big Phillies fan and they just came back [04:27] thanks that helps, I probably should have been looking at that all along [07:58] cjwatson: if you have a moment, can you look at bug 150872. I suspect I'm missing something fairly obvious in my patch, but I'm not sure what. [08:04] evand: So, I'm trying to install on the Jax. First error was: "ClockSetup failed with code 4", and then I tried again and got: "TimezoneApply with code 1" [08:04] Noting your comment in the bug, but the code in question adds $device /cdrom udf,iso9660 user,noauto... so I'm still not sure what would be wrong with the patch. [08:04] hrmm [08:05] StevenK: Not entirely sure just based on that. Can you stick set -x in the right places and post full logs to a bug report? [09:05] Hm. My set -x hackage didn't work as intended [09:15] StevenK, Does syslog have anything interesting? [09:16] Ah ha! [09:16] cp: writing '/target/etc/localtime': No space left on device [09:16] * StevenK kicks it [09:17] evand: It's not a bug, it's running out of space [09:17] * StevenK shuts the damn thing down [09:17] * davmor2 lends StevenK his big hammer less painful on the foot [09:17] Hehe [09:44] cilkay: all the lines that begin "d-i console-setup console-setup/..." or "d-i console-terminus console-terminus/..." are completely broken and you should remove them (as in, they aren't in the right format for a preseed file and you shouldn't need them anyway) [09:45] cilkay: the rest rings a bell, I'm just trying to find the relevant bug ... [09:48] cilkay: try deleting console-setup/variantcode too? I'm wondering if having the empty string there is confusing console-setup [09:51] evand: I commented on 150872 [13:51] evand: could you fix the way the timezone widget continues to scroll even after the mouse exits the widget for 8.10, please? [13:51] we don't have a lot of time left so I'd like that to land [13:56] partman-ext2r0: cjwatson * r786 ubuntu/ (check.d/nomountpoint_ext2r0 debian/changelog): [13:56] partman-ext2r0: * check.d/nomountpoint_ext2r0: [13:56] partman-ext2r0: - Make $RET look for a boolean value (thanks, Nicolas Valcárcel; [13:56] partman-ext2r0: LP: #256459). [14:01] cjwatson: will do [14:06] partman-ext2r0: cjwatson * r787 ubuntu/debian/control: Maintainer and Vcs-Bzr for Ubuntu [14:24] partman-ext2r0: cjwatson * r788 ubuntu/debian/changelog: releasing version 1.15ubuntu1 [14:53] cjwatson: to follow up on sources.list.apt-setup from yesterday, I tried: 1. setting my proxy in mirror/http/proxy; 2. setting my local repo in local0; 3. setting archive.ubuntu.com in local1; [14:54] cjwatson: the problem is that I'm not getting a Broken package error relating to gimp, probably because it's getting some packages from mirror/http/hostname and some from local1: http://people.ubuntu.com/~cr3/syslog [14:54] cr3: why aren't you setting mirror/http/hostname? [14:54] that seems like the obvious approach; putting the main mirror in local doesn't make sense [14:55] cjwatson: sorry, mirror/http/hostname is set to my repo taken from the alternate image [14:55] cjwatson: whereas local0 is set to a small repo of my own packages [14:55] cjwatson: previously, local1 and the proxy was being set by a custom udeb which created sources.list.apt-setup and a proxy file under /target/etc/apt/apt.conf.d [14:56] cjwatson: the repo taken from the alternate image is indeed the main mirror from which I want to grab packages [14:56] I seem to be being unclear somehow. I think you should set your snapshot of archive.ubuntu.com in mirror/http/hostname and mirror/http/directory, not in local. [14:56] cjwatson: I'll upload the preseed too, one sec [14:57] I thought that was what I'd said yesterday. [14:57] apt-setup/local* is for things like your small repository of your own packages, not for the primary mirror that goes first in sources.list [14:58] cjwatson: I don't have a snapshot of archive.ubuntu.com though, I only have a copy of the alternate image in mirror/http/hostname which is just a subset of archive.ubuntu.com. so, that's why I set archive.ubuntu.com as local1 [14:58] you should put it in mirror/http/hostname anyway [14:59] cjwatson: here's my preseed: http://people.ubuntu.com/~cr3/preseed.txt [14:59] do you ever actually want to retrieve any packages from archive.ubuntu.com? [14:59] it seems like asking for trouble in your setup [15:00] cjwatson: yes, I need some packages from archive.ubuntu.com which are dependencies for my tests. however, I just want those dependencies from archive.ubuntu.com, the rest should come from the alternate image [15:00] where do you install those dependencies? [15:00] cjwatson: pkgsel/include typically refers to my own package, which is available on local0 and which specifies dependencies in the package itself [15:01] cr3: does that package depend on things that need to be retrieved from archive.ubuntu.com? [15:01] cjwatson: yes, because they're not available on the alternate image [15:02] I see. [15:03] cr3: The only way apt supports doing this reliably is using /etc/apt/preferences (which the installer has no built-in means of writing although it might be possible in a post-base-installer script). However, in this case I think it probably makes more sense just to append bits to sources.list in late_command and install the package there. [15:03] cjwatson: before, using that sources.list.apt-setup trick worked to cover my special case. however, since this was an undocumented feature, I would like to take this opportunity to reexplore my setup and perhaps come up with something more standard. [15:03] cr3: You'd have had occasional problems in the old setup too if skew got large enough. [15:04] cjwatson: I might've has sporadic problems indeed, I will definately gain from having a more stable solution in the long run [15:05] cjwatson: ok, so I should handle everything in the post-install. but, I have to be careful, because grub is installed afterwards [15:05] late_command is done after grub [15:06] cjwatson: ah, but grub is done after sources.list.apt-setup is copied over, right? that might be why I'm confused [15:07] cjwatson: I'll have a quick look at /etc/apt/preferences before jumping into the late_command solution [15:08] correct [15:10] cjwatson: where can I find documentation or examples for /etc/apt/preferences? I tried looking in busybox during installation, there's no /etc/apt/preferences nor under /target. I also tried grepping the installation-guide-i386 package on Hardy, but not there either. [15:12] apt_preferences(5) === nxvl_ is now known as nxvl [15:28] cjwatson: ok, so I'm thinking that a late_command script will be easier to maintain and more flexible than using /etc/apt/preferences. so, what's the difference between using in-target and chroot /target? [15:32] cjwatson: I think the difference is that in-target logs properly to /var/log/syslog [15:35] cr3, It also takes care of some other accounting for you : it's essentially a convenience script to make your life easy. [15:35] cr3: the main relevant difference is that in-target sets up debconf interaction [15:35] oh, and mounts /proc [15:35] start with in-target but depending on exactly what you're doing you may find that you need to drop back to chroot /target [17:03] I'm trying to create a custom ubuntu .iso but when I boot the mkisofs-created image in Vbox it errors on finding any of the kernels. [17:03] basically I'm trying to create a customized boot.iso with both 32 and 64bit netinstall kernel options at boot. [17:06] currently I have all kernels and initrd in the root folder of the .iso, pointed to by the isolinux.cfg. [17:07] Zelut, Does it fail to boot, or fail to detect a valid kernel for install? [17:18] it says it can't find kernel [17:18] no matter the boot option I select from my custom isolinux.cfg / boot.txt menu it can't find the file. === cjwatson_ is now known as cjwatson [17:23] Zelut, Hrm. That's an isolinux thing, about which I know very little. Sorry. [17:43] does the .iso not allow for renaming the kernel/initrd for some reason? [18:32] when I do a network install, I have debconf/priority=critical as a kernel parameter. however, that doesn't seem to be passed when running in-target apt-get install -y --force-yes in the late_command [18:34] ubiquity: evand * r2898 ubiquity/ (debian/changelog ubiquity/zoommap.py): [18:34] ubiquity: Stop scrolling the timezone map once the mouse is outside its [18:34] ubiquity: boundaries (LP: #251231). [19:44] cjwatson: All that "d-i console-setup console-setup" stuff I put in my preseed file was because the "simplified" file as per the preseed example file provided in the docs did not work. I had the same issues. Once I went through a manual "dpkg-reconfigure console-setup", I paid attention to the questions and answers and found the "d-i console-setup console-setup/" strings in the file created using debconf- [19:44] get-selections. [19:45] I'm at wit's end with preseed installation. [19:45] It seems much more complicated than kickstart. [19:46] I tried using a GUI tool to generate a kickstart file too but when it came to the package selection section, there were no packages to select. [19:48] In case this has fallen off your scrollback, I asked a question on the forum and the Ubuntu Users list about preseeding dpkg-reconfigure console-setup but I didn't get any responses so I suspect it was the wrong place to ask. [19:48] Here is the forum post: http://ubuntuforums.org/showthread.php?t=943668 [20:18] ubiquity: evand * r2899 ubiquity/ (debian/changelog ubiquity/zoommap.py): [20:18] ubiquity: Iterate through a list of nearby timezones on click, rather than [20:18] ubiquity: selecting the absolute closest timezone to the pointer. [20:46] ubiquity: cjwatson * r2900 ubiquity/ (debian/changelog ubiquity/i18n.py): Fix translation of Quit button (LP: #277451). [20:47] evand: are you planning a ubiquity upload today? [20:48] cjwatson: yes, I was going to wait until later tonight though. [21:00] ok this is driving me nuts. I've downloaded a second linux/initrd.gz and added it to the root of my mini.iso. When I select that option at boot it can't find it. [21:01] permissions seem to be the same. ownership is the same. isolinux.cfg has been updated, using the same path syntax as the original.. [21:10] Hi. Can anyone tell me if there's any plans to update the GUI in the installer? [21:11] MadsRH: in what way? [21:11] oh, you were the person asking about slideshow the other day weren't you? [21:12] Hi cjwatson - Yes, I was :-) [21:12] I've been looking in the wiki and the blueprintes, but there's no info about that [21:13] It's been deferred. [21:13] MadsRH: the ubiquity-visual-refresh spec is the current plan of record, but it has largely been crowded out by other more urgent goals [21:13] https://blueprints.launchpad.net/ubuntu/+spec/ubiquity-visual-refresh [21:14] thanks === ago_ is now known as xivulon_ [21:32] So I read the spec for the ubiquity-visual-refresh and it seems like the TODO task would be the next step here. I really want to see this ubiquity-visual-refresh happen and want to help out (if I may, as I'm not part of the installer-team). [21:33] hmm, something does not work in partman/wubi, I am thrown out with "No filesystem is defined" [21:33] * xivulon_ investigating [21:33] "No root file system is defined" [21:34] MadsRH: The mockup is done, it's just buried under a pile of papers on my desk. [21:34] The next step would be to investigate what format to use (animation needed) and how to best overlay translated text on it. [21:35] evand -> Okay, so I can't find in on the web (wiki, mailinglist...)? === cjwatson_ is now known as cjwatson [21:35] MadsRH: No, it was hand drawn at UDS. Let me see if I can find it, at which point I'll scan it and put it on that wiki page. [21:37] Evand -> Great :-) you say hand drawn. Perhaps I could do a mockup of it? But that sounds really good, thanks [21:44] That would be very helpful as my drawing capabilities, as evidenced by the notes I'm about to upload, are quite poor. [21:44] ;-) [21:53] hrm, hopefully he'll notice that I updated the wiki page. [21:53] My notes are not entirely helpful, but I suspect if he has questions he'll be back. [22:23] Is there a known issue with Daily-LiveCD crashing on Manual Partitioning? I just tried an install using [22:24] today's CD, and got Partman failed with exit code 141 [22:26] charlie-tca: yes [22:26] https://launchpad.net/ubuntu/+bug/282756 [22:27] Thanks [22:27] it's fixed in tree, just awaiting a new ubiquity upload, which I'll be taking care of in a few hours. [22:27] Okay, I'll try this again in about 12 hours then [22:55] evand: hey there, you still around? [22:55] TheMuso: or you, perhaps? [22:55] evand: TheMuso I'm hitting something strange in grub-installer [22:55] evand: TheMuso: line 585 looks like: [22:55] kirkland: Sure, whats up? [22:55] if [ -z "$frdisk" ]; then [22:55] TheMuso: hey! [22:56] TheMuso: okay, so running grub-installer through set -x, I'm seeing that frdisk=found [22:56] TheMuso: how and where is that getting set? [22:56] TheMuso: specifically, i'm working with virtio devices, eg, /dev/vda [22:56] kirkland: That gets set if the boot device is either mdadm, sataraid, or multipath if memory serves. [22:57] TheMuso: yup, but it's also getting set when the boot device is vda, and that's breaking me :-) [22:57] TheMuso: I'm trying to find the faulty logic, and I'm thinking it might not be in grub-installer, somehow [22:57] kirkland: Right, well you will have to look and find what sets it for virtio devices, by following the execution path of the script. [22:58] TheMuso: set -x output: http://pastebin.ubuntu.com/57617/ [22:59] TheMuso: I'm bothered that that doesn't show any frdisk=... [23:00] kirkland: oh so you want frdisk to be set? You may have to put another if stanza in to do that. [23:00] TheMuso: ah, okay, that's easy enough to do [23:00] TheMuso: below mdadm or some such [23:01] kirkland: I guess so. [23:35] TheMuso: cool, i have a fix, tested, works [23:35] TheMuso: would you mind having a look and sponsoring, so that it makes it into the next daily build? [23:35] kirkland: Sure. [23:36] * TheMuso installs cia-clients again since he is using a fresh install... [23:37] partman-base: TheMuso * r113 ubuntu/debian/changelog: releasing version 121ubuntu8 === ago_ is now known as xivulon_ [23:40] hw-detect: TheMuso * r88 ubuntu/ (debian/changelog debian/disk-detect.templates disk-detect.sh): [23:40] hw-detect: disk-detect.sh, debian/disk-detect.templates: If dmraid arrays are found, [23:40] hw-detect: ask the user if they want to use them. (LP: #279288) [23:40] TheMuso: https://bugs.edge.launchpad.net/ubuntu/+source/grub-installer/+bug/281492 [23:40] TheMuso: patch attached at the bottom, as well a bzr branch [23:41] kirkland: Thanks, will attend to it as soon as I upload these two packages. [23:41] TheMuso: cool [23:41] cjwatson, evand, I think that /lib/partman/check.d/proper_mountpoints creates issues with wubi [23:42] hw-detect: TheMuso * r89 ubuntu/debian/changelog: releasing version 1.63ubuntu4 [23:43] xivulon_: are you sure about that? It's existed as is since March. [23:43] evand doing some tests now, installation stops there with "No root file system is defined" [23:43] well, maybe it isn't. [23:44] it's a warning message after all, sometimes its job is to fail [23:44] I suspect it's failing elsewhere and just bubbling up there. [23:44] the no_root check is really, really simple. I agree with Evan [23:45] yep looks like that, looks for / in fstab.d [23:45] 8/c [23:50] Hi, when using usb-creator with a persistence file does it save all changes to the file system from the live image, or is there a certain location to save to? [23:50] or am I missing the intent? or the limits of what it can do? [23:50] have that both in vm and real hardware [23:54] james_w: it's broken so it does neither at the moment [23:54] but it will be all changes to the filesystem [23:54] casper affords two options, as I understand it [23:54] one is to do the entire filesystem, the other is to do /home [23:55] usb-creator opts for the former [23:55] ah, good to know it doesn't work :-) [23:55] I thought I was being stupid [23:55] https://bugs.launchpad.net/ubuntu/+source/casper/+bug/274076 [23:55] thanks [23:58] james_w: please do file any bugs you find in it without persistence enabled. I am keen on getting as much testing as possible before release. [23:58] with usb-creator? [23:58] ja [23:58] yeah, I've been having trouble with it writing dud images, I can't work out what the cause is though [23:59] it's something to do with formatting [23:59] boot flag not set? [23:59] It could be, when I choose USB from the boot menu I get nothing more than a flashing cursor