[07:06] usb-creator: evand * r46 usb-creator/ (debian/changelog usbcreator/backend.py): [07:06] usb-creator: Fix crash when the shutdown function would get called before the install [07:06] usb-creator: process began (LP: #277869). [07:17] usb-creator: evand * r47 usb-creator/ (4 files in 3 dirs): Added an icon (taken from Ubiquity until we have an icon of our own). [08:48] evand, cjwatson: netboot mini.iso still stop at the boot: prompt rather than the selector screen. [08:48] usb-creator: evand * r48 usb-creator/debian/changelog: Added LP bug reference. [08:49] new kernel? [09:18] I don't think it would be, but can anyone see anything especially evil with doing the following as part of usb-creator's bootloader installation: dd if=/usr/lib/syslinux/mbr.bin of=/dev/sdb bs=446 count=1 conv=sync [09:19] in addition to calling syslinux regularly [09:20] I'm trying to mitigate issues people are having with No Bootloader Found messages and the possibility that users could have another bootloader already installed to the MBR [09:21] where /dev/sdb is the parent device of the target install partition. [09:28] evand, It should be safe under the assumption that people don't already have a bootloader on their USB stick. [09:29] that they don't want to keep using, of course [09:29] but right [09:33] evand, Well, given the effort to ask users to manually delete stuff when using usb-creator twice on the same key, it's probably worth thinking about a more robust solution for jaunty, but for now, that seems a reasonable hack. [09:34] What do you mean by delete twice on the same key? [09:35] Last time I used usb-creator, if I wanted to put a different image on the same 1GB key, I was given a nautilus window to delete stuff to make space for the image. That's what I mean, although if you changed it, it becomes moot. [09:36] oh, of course. It doesn't currently factor in what files its going to delete as part of the free space calculation. That's a glaring bug that I need to fix. [09:36] persia: However, does this connect somehow to the bootloader discussion? That's where I'm getting confused. [09:37] Just that because the tool is careful about the contents of the data partition, being similarly careful about the contents of the boot partition seems reasonable. [09:38] Mind you, that's not something worth fixing now, as it's a string change, and would break translations, etc. [09:39] persia: ah, indeed and noted. Thanks for the advice. [11:00] cjwatson, i didnt get the help from ubuntu for getting the output from grub prompt .. please if you know tell me [11:13] cjwatson, i didnt get the help from #ubuntu for getting the output of grub prompt to any file .. tell me if you know the answer [11:16] ganes, Repeating yourself here is very unlikely to make any difference in the response you get, other than perhaps comments like this one discouraging repetition. [11:22] persia, ok [12:27] cjwatson: thanks again :-) it works [12:27] cool [12:27] sorry it's so kludgy, it gets a bit easier in intrepid (with the introduction of partman/early_command) but we could still do with a more direct syntax [12:29] correctly wrapping lines and a little comment above make it really understandable [12:55] cjwatson, can you answer me [12:55] no, sorry. [12:55] this is a developer channel, not for user help [12:57] I'm happy to help people out when it stands a good chance of the installer being improved as a result, but you're just repeating questions over and over again. I've already given you the basic tools that may help you get where you need to go, but beyond that I'm afraid you're going to have to sort it out yourself [12:58] cjwatson, ok , you gave command with that i tried eventhough it is not solving my problem [12:58] ok, I'm sorry but I am not available to help you any further [12:59] please do not ask further questions about redirecting output from grub to a file in this channel [12:59] ganes, i have customised the hardy ubuntu filesystem but the ubiquity is not working [13:00] ganes, before getting the password .. it is searching & stopping there itself why [13:01] cjwatson, the above query is for you .. sorry [13:01] cjwatson, little bit tense [13:01] ganes: this is a developer channel. [13:02] if you have customised it and it broke, it is up to you to figure out which of your customisations was responsible. [13:02] I have told you this before. [13:02] cjwatson, but i have not changed anything in ubiquity [13:03] cjwatson, that only surprising me [13:03] You produced a customised image. Your image is broken where the original is not. Therefore your customisations were responsible. [13:03] This is really simple logic. [13:03] cjwatson: netboot mini.iso is still stopping on that first boot: prompt rather than the install selection screen. [13:03] ganes: I accept that you do not believe you have changed anything in ubiquity, but I cannot help with everyone's customisations of Ubuntu. I have enough to do with Ubuntu itself. [13:04] ganes: Please do not ask us any further questions about your customised image unless you're coming with suggested patches. [13:04] cjwatson, it is booting every application is running perfectly execpt this ubiquity.. i shown to many peoples .. but there is no solution [13:04] I'm sorry to have to say that but you have been badgering us for weeks. [13:04] davmor2: I know, haven't figured out that bug yet :( [13:05] cjwatson: okay np's just letting you know :) [13:05] cjwatson, ok thanks for your kindness [13:05] ganes: the installer is quite sensitive to some relatively small details of CD construction (e.g. the presence of .disk, which most applications won't care about). You have to get it exactly right. === davmor2 is now known as davmor2_lunch [13:06] cjwatson, you said this before itself .. i did eventhough [13:06] We need to be able to use this channel for development. I must ask you to stop asking customisation questions here. [13:07] cjwatson, anyhow i am happy of seeing my customised image .. following your ubuntu docs [13:07] cjwatson, ok stopping [13:07] thank you [13:31] fyi my issue with X was due to having left nolapic boot arg [13:39] persia: decided to wander in at least [13:39] :) [13:40] evand: Im told you may be interested in having a peek at bug 286950 [13:40] err, no ubottu clone in here? have to fix that... [13:41] jussi01, Don't put a bug bot in here, please. [13:41] persia: ok. [13:42] 286950 is Forbidden [13:42] so a bug bot would make little difference, as it wouldn't be able to see it [13:43] cjwatson: but it would give a url... [13:43] I'd be OK with a bug bot in here [13:43] bugs.launchpad.net/ [13:43] bugs.launchpad.net/bugs/ you mean [13:44] Er, yes [13:44] To be perfectly honest, I have a Firefox shortcut 'lpbug ' and it just fixes it [13:44] jussi01: install gksu [13:44] cjwatson: if you want her gone, ping me [13:44] should be fine, thanks [13:48] cjwatson: seems to be working fine now. perhaps we need to add a dep to the package? [13:48] should fix the privilege escalation instead. all the world doesn't use gksu :) [13:48] but a dependency would be good for now, yes [13:48] anyway, wait for evand to show up :) [13:48] ok, (and yeah, Id love to see it using kdesu ) [13:50] that would make more sense if it had a KDE frontend, don't you think? :) [13:50] cjwatson: exactly my point... ;) [13:52] anyway, I hang around for evand if he needs any info. === davmor2_lunch is now known as davmor2 [14:05] usb-creator: evand * r49 usb-creator/debian/ (changelog control): Add dependency on gksu. Thanks Colin Watson (LP: #286950). [14:06] jussi01: ^ [15:15] cjwatson: Do you remember if the installer does anything special w.r.t. to /var/run/network ? I'm trying to work out why https://bugs.edge.launchpad.net/vmbuilder/+bug/276365 is necessary. [15:15] Launchpad bug 276365 in vmbuilder "/var/run/network is not created" [Undecided,Fix committed] [15:15] I don't see how it could, that would be cleared on boot [15:16] cjwatson: Exactly. [15:16] /etc/init.d/loopback creates /var/run/network [15:16] I don't suppose you've removed ifupdown? [15:16] Well, it makes a bit of a difference.. [15:16] since that's where that init script lives [15:17] if /var/run/network exists on the filesystem, it'll be availabale before /var/run is mounted and /var/run/network is created, but I don't see why it's necessary. [15:17] cjwatson: Let me check.. [15:18] 14:11:27 < nijaba> soren: it only occurs when /var is on a separate partition/disk [15:18] 14:12:14 < nijaba> soren: when the vm first start, there is a process that tries to create /var/run/network during first boot, but fails as it tries to do so before /var is mounted [15:18] 14:12:31 < nijaba> soren: subsequently all network operations fail [15:19] Ah, so it might jsut be that /var/run needs to exist on the root file system.. [15:19] That still doesn't explain why nijaba's patch fixes anything, though. [15:19] now, there is some careful stuff in the installer for /var/run [15:20] evand: thanks :) [15:20] cjwatson: Don't spend any time on it. I'll work it out. I just wanted to hear if you knew that you did something special to handle it in the installer. [15:20] cjwatson: Oh, really? [15:20] see partman-target/finish.d/mount_partitions [15:20] if /var is a separate filesystem then you need to arrange to create /var/run and /var/lock on the root filesystem *before* /var is mounted [15:21] cjwatson: Got it, thanks. [15:21] see bug 44142 [15:21] Launchpad bug 44142 in partman-target "Doesn't create /var/run or /var/lock directories underneath /var mountpoint" [Critical,Fix released] https://launchpad.net/bugs/44142 [15:55] I've reproduced the persistence shutdown errors on another machine, but I can't get a different USB stick to boot. I can buy another if you think it could be the stick's fault. [15:57] james_w: where does the boot fail? [15:57] evand: no it's the (aufs?) errors on shutdown that I'm seeing [15:58] james_w: indeed, I was just curious as to why you couldn't get the other USB stick to boot. [15:58] james_w: Can you create a bug report with your findings thusfar so we can track this? [15:58] If you have not done so already [15:58] oh, the BIOS just doesn't pick it up for some reason, I'll try another machine [15:59] it is a lot larger, so I don't know if it's just too large for that machine to use [15:59] james_w: Created using usb-creator? It's a bug, install-mbr /dev/$device should fix that for you. [15:59] that'll be it [15:59] this machine is a bit more verbose [16:00] /dev/sdx not /dev/sdx1 I assume? [16:00] correct [16:01] would you like the bug on casper? [16:01] indeed, thanks [16:09] yep, second key (same manufacturer) shows the same symptoms. [16:17] bug 287027 [16:17] Launchpad bug 287027 in casper "persistence leads to filesystem errors on shutdown" [Undecided,New] https://launchpad.net/bugs/287027 [16:21] thanks [16:24] xivulon, http://pastebin.com/f557ba44f is the log [16:25] xivulon, oops that's not it, let me find the right one [16:25] xivulon, http://pastebin.com/f67afbe64 is the right one. it only happens if you are connected to a network, but NOT connected to the internet [16:32] tasksel: cjwatson * r1384 ubuntu/ (debian/changelog tasksel.pl): [16:32] tasksel: Tell apt to install Recommends, just in case tasks are incomplete in [16:32] tasksel: this regard (e.g. a task containing a metapackage that manually depends [16:32] tasksel: on a metapackage for another task; LP: #276317). [17:05] might there be an appropriate place to set gconf variables for a particular user during the installation process? [17:07] worst case, I'll just do: in-target sudo -u user gconftool-2 ... [17:08] that's about the best available === davmor2 is now known as davmor2_tea === davmor2_tea is now known as davmor2 [18:51] xivulon: wubi removal bar works well now :) [19:02] cjwatson: http://pastebin.ubuntu.com/60623/ [19:03] cjwatson: i just confirmed a report that installs to raid0 and raid5 devices are broken [19:03] cjwatson: i'm verfied that pastebin'd fix for the raid0 case [19:03] cjwatson: i'm testing raid5 and regression testing raid1 now [19:36] davmor2 cool [19:36] superm1 the last log seems ok [19:38] davmor2, I got stacked yesterday running the system cleanup utility, didn't dig though [19:38] superm1, let me find the "wrong" log [19:40] hmm all logs look good to me [19:41] xivulon: system cleaner basically says there is no relatime option in the host fstab line [19:44] xivulon, yeah its just at the "reboot" screen that wubi's exe crashes [19:45] xivulon, so it did finish "successfully" doing all the other steps [19:45] you should be able to reproduce it if you set up virtualbox with networking but then disable your wireless or unplug your cat5 [19:46] i can reliably reproduce it by doing that [20:02] superm1 will try that [20:04] by the way I thought relatime was the default... having the old atime with no caching and syncio in ntfs isn't going to be to be too nice... [20:07] well in fact that should be mostly irrelevant [20:08] but is it now usual to have explicit relatime in fstab? [20:10] xivulon: might be worth talking to liw about. He's system-cleaner's main author. [20:12] evand, will do [20:14] evand liw suggests to add relatime to wubi, too late to do that? the change is simple enough on my side [20:15] hmm maybe not, the / line is set in partman-auto [20:15] was thinking about /host [20:16] xivulon: I am fairly certain that the only kind of change that can get in now is something very, very critical. That said, slangasek and #ubuntu-release would be able to give you a much better assessment. [20:16] let me quickly find the relevant code [20:17] FYI, slangasek stepped out for a bit, but I imagine he'd answer such a question promptly when he returns. [20:36] evand: do you know anything about open_infifo? [20:36] evand: /lib/partman/fstab.d/ext3 is hanging there [20:37] cannot create /var/lib/partman/infifo [20:37] that file exists [20:37] has: [20:37] PARTITIONS =dev=md0 [20:37] but that's all [20:38] i have multiple md devices in this installation [20:38] md0 is raid1 [20:38] md1 is raid5 [20:38] cjwatson: ^ [20:38] hrm, not sure. Can I see a full set of logs (syslog and partman)? [20:40] evand: http://pastebin.com/f7d917a80 [20:40] evand: http://pastebin.com/f5b0feafc [20:42] evand: btw... i bought pair of lizard-skin cowboy boots last weekend ;-) found a thrift store with some sweet deals on boots [20:43] nice! [20:44] kirkland: Can you repaste the second log to http://paste.ubuntu.com ? Pastebin.com seems to be having issues. [20:44] cjwatson: evand: I posted a one-line patch for https://bugs.edge.launchpad.net/ubuntu/+source/mdadm/+bug/287112 [20:44] Launchpad bug 287112 in mdadm "intrepid installer hangs when installing to raid0 or raid5" [High,In progress] [20:44] evand: http://pastebin.ubuntu.com/60655/ [20:45] evand: be back in 10 minutes, need to grab a coffee [20:46] given your comment in the bug report about testing it against !raid1 and raid, I'd be happy to sponsor that. [20:48] hrm, do we not keep this in a VCS? [20:51] talking of raid0-1 I haven't tested wubi on that, could someone with software raids give it a go? [20:52] indeed, looks like partman-auto-loop's fault [20:55] it's my understanding that you cannot call close_dialog until you've read it all [20:55] in fstab.d/hostboot that is [20:55] might be a red herring though [20:56] evand are you talking about fstab mount options? [20:57] I think that refers to the / line not sure adding mount options to bind mounts is a good idea (probably they would be ignored) [20:57] yes, fstab.d/hostboot [20:57] kirkland: can you stick a set -x at the top of that and give it another go [20:58] evand: all of my raid testing has been in vm's [20:58] evand, hostboot generates the /boot line in fstab, the / line is generated in some other partman library (forgot which) and having to use IE6+LP to search the code is less than optimal :( [20:58] evand: at the top of what? [20:58] /lib/partman/fstab.d/hostboot [20:59] yep that generates boot, there is one other script that generates the line for / [20:59] evand: and "give another go" means running it from the command line, or restarting the install [20:59] restarting the install please [21:00] evand: okay, it takes me a while to do all of that partitioning, gimme a few minutes ;-) [21:00] no worries, I'm here for another hour [21:00] and will be back after that around 7-8 [21:02] kirkland: I've sponsored your mdadm upload, but I'll leave it to you to take the case to -release for why it should be accepted and the CDs regenerated again. [21:03] xivulon: sorry about that, I misread you the first time around [21:04] evand: thanks [21:13] evand: okay, i'm at the point in the installation just before the error [21:14] ok [21:14] evand: i added -x to /lib/partman/fstab.d/hostboot [21:14] set -x, right [21:15] evand: and manually edited 07root_on_raid to add my fix [21:15] ok [21:20] evand: i'll let the raid finish resyncing and then i'll pastebin the logs [21:20] evand: what do you want? syslog and partman again? [21:21] ja [21:23] evand: http://pastebin.ubuntu.com/60670/ [21:23] evand: http://pastebin.ubuntu.com/60671/ [21:35] looking [21:42] evand: thoughts? [21:45] not sure, nothing is standing out [21:58] kirkland: so only RAID1 can be degraded? [21:58] cjwatson: well, RAID0 clearly cannot be degraded [21:58] cjwatson: there's data missing in that case [21:58] cjwatson: RAID5 should be able to be degraded [22:09] cjwatson: but /boot can't be on a RAID5 device [22:10] oh, right [22:10] sure, patch seems fine then [22:10] check that it does the right thing with non-RAID devices obviously (i.e. skips them ...) [22:11] evand: hmm, yes, fstab.d/hostboot is on crack [22:39] indeed [22:40] evand: Did you get around to filing that bug against dmraid? [22:41] evand: something like http://paste.ubuntu.com/60697/ maybe? [22:41] themuso: not yet. will do when i get home, but the good news is that dmraid definitely works on my borrowed Dell system [22:42] cjwatson: no idea how to open a link from the ssh app in my blackberry, but ill take a look when i get back (roughly 1 hr) [22:43] unless youre happy to just commit it [22:45] hi, i tried to install kubuntu per netboot+netinstall on a encrypted lvm partition (tried several times manualy and guided paritioning). grub failed on reboot. is this the right channel to talk about it? [22:45] evand: Ok. [22:51] AmenophisIII: what release? [22:51] 8.10 beta [22:51] AmenophisIII: we fixed a problem very like that just yesterday [22:51] AmenophisIII: are you seeing /dev/dm-0 in the root= line in grub? [22:52] iirc yes [22:52] AmenophisIII: yeah, try again with a current archive [22:52] tried over the weekend... ended up with debian lenny :) [22:52] AmenophisIII: this was ultimately due to dmsetup-udeb missing a set of udev rules [22:52] devmapper (2:1.02.27-3ubuntu2) intrepid; urgency=low [22:52] * Actually install udev rules in dmsetup-udeb, as was the case in hardy. [22:52] -- Colin Watson Mon, 20 Oct 2008 21:10:45 +0100 [22:53] thanks [22:53] I haven't tested with encryption yet, but it was broken without encryption too and I've verified that that case is fixed [22:53] and yes, this is the right place :) [23:00] just to be sure... http://archive.ubuntu.com/ubuntu/dists/intrepid/main/installer-amd64/current/images/netboot/ these are the daily built ones.. with that patch included right? [23:01] ubiquity: cjwatson * r2916 ubiquity/debian/changelog: releasing version 1.10.7 [23:01] im not familiar with ubuntu stuff at all, sorry :) [23:02] ubiquity: cjwatson * r2917 ubiquity/ (debian/changelog gui/qt/liveinstaller.ui): restore Jonathan's change [23:02] AmenophisIII: err, yes and no [23:03] AmenophisIII: the images in installer-*/current/ only contain the very core of the installer; it fetches the rest of itself from the archive [23:03] AmenophisIII: so the patch is not included there, but only because the affected code isn't there either. If you simply rerun it using a current archive then it will fetch code including the patch [23:03] roger [23:06] last question (hopefully :) the guided partitioning lacks functionallity: if there is an existing partition you dont want to touch, you geht asked, if you wanna use the remaining biggest block of free space or something... but you cant do lvm+encryption then [23:07] it just makes a normal ext3 / and swap [23:07] you can do everything the automatic partitioner does manually if you prefer, it's just more work [23:08] that said I'm not sure why automatic lvm+crypto shouldn't work in that case [23:08] ok np for me... i just wanted to mention that, since im already here :) [23:09] oh, it's a separate path - could you file that as a bug on partman-auto-crypto please? [23:09] https://bugs.launchpad.net/ubuntu/+source/partman-auto-crypto/+filebug [23:10] it's not done right now because it would require converting that into a two-level menu, I think [23:10] *registering [23:22] https://bugs.launchpad.net/ubuntu/+source/partman-auto-crypto/+bug/287255 is that enough/what you were thinking about? [23:22] Launchpad bug 287255 in partman-auto-crypto "lack of options for remaining free block" [Undecided,New] [23:23] yeah [23:23] thanks [23:46] cjwatson: obviously looks reasonable, but I don't have a Windows partition to test it with wubi (VMWare doesn't like 2.6.27). kirkland can you test that it fixes the issue for your setup? [23:47] evand: hmm, what are you asking me to test?