[07:01] got a weird issue with the installer, i get to the prepare partition section of the installation but no partitions come up, i used sudo blkid it see's my hdd, i can mount the hard drive i want to install ubuntu to, i can do a sudo mount -t vfat /dev/sda1 /mnt to my hdd and it mounts but partman [07:02] (after some reading i discovered this is what the installer uses to find the partitions) can't seem to figure out i have a hdd [07:32] Shadowmancer: if you can wait around for a couple of hours, I'd be interested in debugging this with you when I get into the office [07:32] ]ok [07:32] that would be alright [08:59] Shadowmancer: ok, so just as a quick check, can you please run 'sudo parted -s /dev/sda print' [08:59] sorry i got to go damn dinner [08:59] i'll try to be back in a couple of hours [08:59] ok [09:06] ubiquity: cjwatson * r4112 ubiquity/ (d-i/manifest debian/changelog): [09:06] ubiquity: Automatic update of included source packages: console-setup [09:06] ubiquity: 1.34ubuntu15, flash-kernel 2.13ubuntu17, hw-detect 1.73ubuntu4, [09:06] ubiquity: migration-assistant 0.6.5, partman-auto 89ubuntu8, partman-partitioning [09:06] ubiquity: 72ubuntu4. [09:10] ubiquity: cjwatson * r4113 ubiquity/debian/changelog: releasing version 2.2.22 [09:32] cjwatson, morning dude [09:38] DASPRiD: hi. I'm going through your video and attempting to match it using some virtual disk images on an external disk. [09:38] may take a while. [09:38] i hope the video helps finding the problem :) [09:39] we'll see [09:49] cjwatson, I uploaded the information you required for bug 568312 [09:49] Launchpad bug 568312 in partman-iscsi "Installer fails to find a target when using authentication" [High,Triaged] https://launchpad.net/bugs/568312 [09:51] ah, thank you [09:52] Apr 26 08:38:35 main-menu[335]: (process:6453): iscsiadm: [09:52] Apr 26 08:38:35 main-menu[335]: (process:6453): cannot make connection to 192.168.101.100:3260 (111) [09:52] networking issue? [09:52] 111 is connection refused, isn't it? [09:53] cjwatson, networking is not the issue, as the same configuration without authentication works perfectly [09:54] understood - but that's a low-level error, not subject to auth/non-auth difference, so I don't quite understand [09:54] I'm also seeing (different) differences between auth/non-auth [09:55] very confusing [10:00] cjwatson, do you want me to run non-auth with "set -x" ? [10:00] (or any other configuration) [10:08] ara: might as well, though I'm debugging something else just at the moment [10:08] but it can't hurt [10:08] DASPRiD: reproduced [10:09] cjwatson, ok, will do and will attach the log to the bug report [10:09] DASPRiD: I think I misunderstood your bug to start with; I didn't realise you were starting out with blank disk images [10:09] heh :) [10:09] and i thought that was clear :) [10:12] so I guess the question is why you're allowed to have that new /boot partition marked as "keep and use existing data" *at all* [10:12] there's no existing data to keep! [10:14] DASPRiD: but you said you could mount /dev/md0 while the partitioner is running [10:14] that doesn't seem consistent with blank disk images? [10:15] cjwatson, uh, did i? [10:15] i did only do what i did in the video [10:16] er, wait, you're not the original bug reporter are you [10:16] * cjwatson kind of wishes that people would file new bugs, and let us mark them as duplicates if they are :-/ [10:18] in the original bug report, it seems clear that /boot already exists [10:18] so are you quite sure that you have the same bug? [10:21] (I'm not sure you *don't* have the same bug either - it just doesn't seem obvious) [10:30] cjwatson, Launchpad interface drives people to mark test as duplicates, rather than filing a new one, maybe that's why [10:31] cjwatson, maybe it should change that to certain packages where the maintainers prefer new bugs (xorg, linux, ubiquity...) [10:37] ara: heh, we were just having that debate in person. my argument is that it makes sense to try to reduce duplication for crash reports, but for anything where the user types in something more descriptive themselves, they should *never* be guided towards possible existing duplicates regardless of package [10:37] to me there's a clear difference between the desirable workflow for crash reports and for descriptive bug reports [10:38] sure, but for packages with little maintenance, maybe a duplicate makes sense [10:39] I wouldn't like to participate in a seahorse bug day (i.e.) where the main activity is marking tons of duplicates of the same bug [10:39] in that case I think that Launchpad interface works OK [10:46] back cjwatson [10:46] :) [10:47] you missed your slot, hang on :) [10:47] but get me that parted output when you can? [10:50] sure thing [10:51] ok so i can't copy and paste the output since the computer i'm working on has no net connection [10:52] but in short it see's the hdd, identifies it as primary, see's the partition table as msdos, its flagged as bootable (i did that with ubuntu) fs is fat32, and each sector size is 512 [10:54] huh, I wasn't expecting that [10:55] ok, I will need some log files from the installer, so can you get a USB stick into it or something? [10:55] specifically I need /var/log/syslog and /var/log/partman [10:55] yeah sure thing [10:57] cjwatson, i'm pretty sure it's the same bug, as me and the reporter experience the same symptoms [10:57] DASPRiD: yeah, you have different situations leading up to it though [10:57] just as a general point, please don't assume that same symptoms => same bug [10:57] I'll leave this as it is for now since it could well be the same thing, but the logic isn't sound that's all [10:58] well, as long as after a fix i can install on two blank disks, i'm fine ;) [10:59] wow i some how killed nautilus :P [10:59] now i need to command line copy [10:59] go me :P [10:59] sorry to hammer on about this, but the problem is that jumping into an existing bug often actively sucks developers' time by having to try to disentangle problems [11:00] anyway, as I say I can reproduce *your* problem, just no idea if it'll deal with the original reporter's bug so am I supposed to close the bug when I fix it or not? :P [11:00] /dev/md0 is definitely missing the ext3 magic number by the time it comes round to try to mount it, but I can see mkfs.ext3 being run [11:01] trawling through the logs to try to figure out what else is eating it in between [11:03] yeah [11:03] well the odd thing with partmans log which i checked already is that it only calls my flash drive [11:03] the one the live disk is loaded into :/ [11:03] let me do the interpretation please :) [11:04] it's not always straightforward to interpret these logs [11:04] so I'd rather just get the log files and look at them [11:04] DASPRiD: hm, unfortunately /dev/md0 is (of course) mkfsed way early, so most of the partitioner's operations come between that and mount :-/ [11:05] yeah ofcourse, would you like the ubiquity debug log? [11:05] sure, /var/log/syslog /var/log/partman /var/log/installer/debug [11:05] that's all I need though, no others [11:08] oh crap, here we go [11:08] /lib/partman/commit.d/50format_ext3: Try to create file system for /var/lib/partman/devices/=dev=md0/0-998244351 [11:08] [...] [11:08] /lib/partman/init.d/49md: IN: NEW_LABEL =dev=md0 loop [11:13] just put them in paste bin? [11:14] sure, whatever [11:14] eventually a bug report but that will do for now [11:15] Hi all, i was about to fill a wishlist-bug in the launchpad for ubiquity, but i figured i would drop by here before to get some comments first - so here it is: from my point of view it would be useful to add a way to save the mbr before installing grub (it's only useful 1 out of 100 install, but when it is it saves tones of time and stress for new users). A great place for that would be an option in the "advanced" menu at step 7 [11:15] in ubiquity that would wrap "dd" and allow the selection of a volume to copy it to (usb disk or whatever) [11:15] arf ... sorry for the long post [11:15] shouldn't be done in ubiquity at all. there's already a bug report on grub2 for it. [11:15] and it's been brought up upstream. there was some contention but it really ought to be fixed there. [11:16] it certainly shouldn't have any UI in ubiquity [11:16] ok thanks for the info [11:17] I can't find the master bug report for it just now, but it's definitely been filed somewhere :) [11:18] cjwatson, no problem ... i should be able to find it :) [11:18] not going to change for lucid now though, sorry [11:18] will try to do something about it for maverick [11:19] I was not thinking about lucid ... i believe this sort of thing to be pre-alpha considerations ... [11:19] http://pastebin.com/YWbA9k5H [11:23] http://pastebin.com/9gDQiSDw [11:23] sorry its taken so long [11:23] i was doing 101 things [11:23] tell me about it ... [11:25] Shadowmancer: can you run 'sudo parted_devices' and show me the output? [11:26] Shadowmancer: also, are you running the installer from the disk you're trying to install to? [11:26] no [11:26] i am using a thumb/flash drive with the installer to install to a sata hdd [11:35] Shadowmancer: ok, how about that 'sudo parted_devices' output then? [11:36] yeah its coming, the damn thing is going weird about the flash drive i used to transport it so i had to reboot it [11:40] http://pastebin.com/fXiZM1qK [11:41] any RAID, fake-RAID, or multipath involved here? [11:41] (that you know of) [11:42] i checked the grub bugtracker ... saving the mbr has been included as a task ( http://savannah.gnu.org/bugs/?func=detailitem&item_id=27058 ) ... but the task is closed ( http://savannah.gnu.org/task/?func=detailitem&item_id=9252 ) because "Users who would know about this option should be able to use `dd' instead." ... i don't get the feeling that reporting this again would achieve anything [11:42] it wouldn't :) [11:43] :) [11:43] there's been disagreement upstream, that closure probably isn't final [11:43] when I have some time I'll bring it up again [11:44] thanks ... that would ease the work of people trying to help users with tatooed laptop [11:46] nope [11:46] its a single Sata hdd [11:46] old thing [11:46] a few years old [11:47] so rather than playing twenty questions, let's have you get a detailed log [11:47] Shadowmancer: please reboot, and get to the live session rather than going straight into the installer [11:47] Shadowmancer: edit /lib/partman/init.d/30parted as root - you'll find a line 'set -e' near the top, change that to 'set -ex' [11:48] Shadowmancer: then start the installer from a terminal by running 'ubiquity -d' [11:48] sure thing [11:48] Shadowmancer: run through the installer until you reproduce the bug, then exit the installer and send me those three log files again [12:01] http://pastebin.com/6C4bheSq [12:10] Shadowmancer: ok, and the other two? [12:11] which other 2? [12:11] 11:48 Shadowmancer: run through the installer until you reproduce the bug, then exit the installer and send me those three log files again [12:12] I asked for three log files - /var/log/syslog, /var/log/partman, and /var/log/installer/debug [12:12] that pastebin item just has the third of those [12:13] ubiquity: evand * r4114 ubiquity/ (debian/changelog ubiquity/components/ubi-timezone.py): Correctly select the detected timezone (LP: #559647). [12:13] Launchpad bug 559647 in ubiquity "[Lucid] "Where are you?" screen doesn't select timezone" [High,Triaged] https://launchpad.net/bugs/559647 [12:16] oh you want the syslog and partman log [12:16] doh missed that [12:16] i'll get them [12:21] DASPRiD: at this point, I can't find any evidence that this ever worked as I think it ought to, so it may not be a regression in lucid which might make it harder for me to get a fix in at this point [12:21] DASPRiD: I think that you should be able to avoid the bug by not setting up /boot until *after* you've done the LVM setup [12:22] DASPRiD: I'm downloading a karmic image at the moment for comparison (and will try hardy if that succeeds) [12:22] hrmmm partman wasn't created [12:22] the partman log didn't get created i mean [12:22] :/ [12:22] never mind that then, just get me syslog [12:23] i am [12:24] okay thanks [12:29] cjwatson, I found a minor issue with iscsi auth, but I wanted to ask you first if you fixed it in version 14. If I enter the wrong user/pass and it gives me auth error, when going back, it loops indefinitely in an error showing "No password was entered" [12:30] yeah, I fixed that one [12:30] * Back up to appropriate previous state after asking [12:30] partman-iscsi/login/empty_password. [12:31] uh, I think so anyway. worth double-checking of course. [12:36] cjwatson, ok, I'll double check when a new ISO is available [12:42] sorry its taking so long the damn thing crashed [12:42] this computer is driving me beserk [12:44] at least not nuts :) [12:46] ubiquity: evand * r4115 ubiquity/ (bin/ubiquity-dm debian/changelog): Don't reboot on failure if we say we're not going to. [12:47] damn, I think this md thing actually stems from a parted change [12:48] so it is actually a lucid one? [12:49] yeah [12:50] pretty sure anyway [13:05] grrrr cjwatson i think i'll call it a night with this problem >.< [13:06] i got to work on other things and all its doing is giving me a headache [13:06] i won't be back at my mums place for a couple of weeks so maybe i can work something better out [13:06] ttyl [14:03] cjwatson, then we clearly need it fixed [14:03] yeah, working on it [14:52] partman-md: cjwatson * r968 ubuntu/ (debian/changelog init.d/md): [14:52] partman-md: Only call NEW_LABEL on MD devices if they're empty, since with current [14:52] partman-md: parted this unconditionally clobbers the superblock (LP: #542210). [14:52] partman-md: cjwatson * r969 ubuntu/ (debian/changelog init.d/md-devices): [14:52] partman-md: Register partman-md/confirm_nooverwrite, associated with the [14:52] partman-md: partman-md/confirm template. [14:52] partman-md: cjwatson * r970 ubuntu/debian/control: Maintainer and Vcs-Bzr for Ubuntu [14:52] partman-md: cjwatson * r971 ubuntu/debian/changelog: releasing version 49ubuntu1 [14:52] Launchpad bug 542210 in debian-installer "/boot destroyed during install even when formatting disabled (alternate installer)" [Undecided,New] https://launchpad.net/bugs/542210 [14:53] partman-lvm: cjwatson * r1280 partman-lvm/ (debian/changelog init.d/lvm-devices): [14:53] partman-lvm: Register partman-lvm/confirm_nooverwrite, associated with the [14:53] partman-lvm: partman-lvm/confirm template. [14:53] partman-lvm: cjwatson * r1281 partman-lvm/debian/control: Maintainer and Vcs-Bzr for Ubuntu [14:54] partman-lvm: cjwatson * r1282 partman-lvm/debian/changelog: releasing version 69ubuntu1 [14:54] \o/ [14:54] partman-crypto: cjwatson * r700 ubuntu/ (debian/changelog init.d/crypto): [14:54] partman-crypto: Register partman-crypto/confirm_nooverwrite, associated with the [14:54] partman-crypto: partman-crypto/confirm template. [14:55] partman-crypto: cjwatson * r701 ubuntu/debian/changelog: releasing version 40ubuntu5 [15:41] If somebody did the work to test etc, could we replace the Live USB tool used in Ubuntu with the one used in Fedora? (which has the advantage of being cross platform among other things) [16:14] ubiquity: evand * r4116 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): [16:14] ubiquity: Always run wrap_fix in its entirety, even if the width or height [16:14] ubiquity: hasn't changed, as we may have switched text direction since the [16:14] ubiquity: last run (LP: #560114). [16:14] Launchpad bug 560114 in ubiquity "Truncated text in Ubiquity after selecting Hebrew" [Undecided,New] https://launchpad.net/bugs/560114 [16:33] ubiquity: evand * r4117 ubiquity/ (debian/changelog gui/gtk/stepKeyboardConf.ui): [16:33] ubiquity: Properly align keyboard_test_label, which was getting cut off in RTL [16:33] ubiquity: languages with the wrap_fix in place. [16:48] the extended_layout stuff might finally land soon: http://mail.gnome.org/archives/gtk-devel-list/2010-April/msg00070.html [17:14] ubiquity: evand * r4118 ubiquity/debian/changelog: releasing version 2.2.23 [18:01] wubi: evand * r189 trunk/ (debian/changelog src/wubi/backends/common/backend.py): [18:01] wubi: str.replace() does not do in place modification of the string [18:01] wubi: (LP: #567969). [18:01] Launchpad bug 567969 in wubi "Wubi's help boot cd options goes to ubiquity automatic" [Undecided,Fix committed] https://launchpad.net/bugs/567969 [18:15] ev, dang, those fixes you made above appear to break some of the third party plugins i've got somehow vertically now. http://imagebin.org/94504 [18:15] fuck [18:16] the regular pages were getting that in oem-config too when you maximized the window [18:17] noted, testing now [18:18] superm1: remind me, is that in a public branch? [18:18] ev, yes [18:18] lp:dell-recovery [18:18] righto [18:20] superm1: does this happen to fix it: http://paste.ubuntu.com/422895/ [18:21] superm1: what page specifically in oem-config? [18:22] ev, the tz page [18:22] i'm trying to get my env back together to try that fix you posted right now [18:22] okay, thanks [18:24] http://imagebin.org/94506 yeah that fixes the one label it looks like [18:24] so i might just have to adjust a bunch of them with that similar change [18:28] hrm, I cannot reproduce the timezone page thing [18:28] cjwatson: is still trying though [18:34] confirmed when I switch to simplified Chinese [18:38] if it's only when the window is maximised, I don't think it's worth a respin at this point [18:42] so far I can only make it happen in zh_CN, but only when maximised, and the page should be intelligible without that anyway [19:18] once i make it happen in zh_CN and switch back to english english is messed up too from what i saw so far [19:18] but yeah zh_CN was the easiest to reproduce it with [20:55] ev, cjwatson so i'm still seeing other core pages get cut off when you switch locales though too; for example german user page: http://imagebin.org/94534 [22:09] superm1: for what it's worth, the german text getting cut off on the user setup page existed before my recent changes, I was just unable to reproduce it [22:09] * ev digs for the bug [22:12] https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/5688851 [22:12] Error: Ubuntu bug 5688851 not found [22:13] bug #568851 [22:13] Launchpad bug 568851 in hexwars "Deleting BITMAPs results in segfault" [Undecided,New] https://launchpad.net/bugs/568851 [22:14] (no) [22:14] whoops [22:14] https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/568885 [22:14] Ubuntu bug 568885 in ubiquity "ubiquity unable to read an option" [Undecided,New] [22:15] Ah :) [22:52] ev, ah. too bad :(