/srv/irclogs.ubuntu.com/2012/04/05/#ubuntu-installer.txt

CIA-32ubiquity: stgraber * r5358 ubiquity/ (d-i/manifest debian/changelog): releasing version 2.10.700:56
stgrabercjwatson: hmm, my 12.04 server just refused to reboot after the last few updates. I got a unaligned pointer error from grub at boot time03:38
stgrabercjwatson: I fixed it by booting into recovery mode and running update-grub + grub-install on both disks but I'm wondering if there's a general problem of grub not updating the mbr record of all disks on upgrade03:38
stgraber(luckily although some 6000km away, that server had kvm access so was fixable ;))03:39
twbI asked this yesterday but IIRC I didn't get an answer:05:48
twbHow do I tell partman to use 1024 units instead of 1000 units?05:49
twb1Mi doesn't work05:49
twbReally confusingly, if I type 2147483648B it claims to allocate 2GB (not GiB), and if I give 1073741824 B it claims to allocate 512.0 MB05:50
twbThis is in precise BTW05:51
twbFuck it, anna-intall parted-udeb and do it by hand (if I can).05:52
twbcjwatson: FYI, "guided partitioning" in precise creates aligned / and /home, but biosgrub is *not* aligned according to parted-udeb's align-check optimal06:00
cjwatsonstgraber: lp:~ubuntu-core-dev/os-prober/ubuntu isn't abandoned, but my branch was unbound and I didn't notice08:16
cjwatsonstgraber: (sorted now)08:16
CIA-32os-prober: cjwatson * r339 ubuntu/os-probes/mounted/x86/20microsoft: Add Windows 8 support08:19
CIA-32os-prober: cjwatson * r340 ubuntu/debian/changelog: releasing version 1.51ubuntu208:20
CIA-32os-prober: cjwatson * r341 ubuntu/os-probes/mounted/x86/20microsoft: Add an alternate mode to 20microsoft where if WINOSDATA is set in the environment, partitions containing the OS are listed instead of partitions containing the boot record.08:20
CIA-32os-prober: cjwatson * r342 ubuntu/debian/changelog: releasing version 1.51ubuntu308:21
cjwatsonstgraber: I've transferred your commits over to that branch now - I'd prefer to use that because being able to merge directly from upstream git is handy08:21
cjwatsonstgraber: grub> what does 'debconf-show grub-pc' say/08:22
cjwatson?08:22
cjwatsonstgraber: os-prober changes look plausible, thanks08:23
cjwatsonstgraber: I've pushed the Windows 8 support part upstream08:25
cjwatsonbdmurray: installation-guide is the right place; I deleted the kickseed task08:28
CIA-32installation-guide: cjwatson * r501 ubuntu/ (build/entities/common.ent debian/changelog): Bump kernelversion to 3.2.08:30
CIA-32installation-guide: cjwatson * r502 ubuntu/ (debian/changelog en/welcome/what-is-ubuntu.xml): Bump GNOME version to 3.4.08:31
CIA-32installation-guide: cjwatson * r503 ubuntu/ (build/entities/common.ent debian/changelog): Bump release version and names for Precise.08:32
CIA-32installation-guide: cjwatson * r504 ubuntu/ (debian/changelog en/install-methods/automatic-install.xml):08:38
CIA-32installation-guide: Add an example Kickstart configuration file, thanks to Ameet Paranjape08:38
CIA-32installation-guide: (LP: #969568).08:38
CIA-32installation-guide: cjwatson * r505 ubuntu/ (build/preseed.pl debian/changelog):08:40
CIA-32installation-guide: Expand &archive-mirror; to archive.ubuntu.com when generating example08:40
CIA-32installation-guide: preseed file (LP: #955607).08:40
cjwatsonreally ought to merge with Debian but this is a start ...08:41
CIA-32installation-guide: cjwatson * r506 ubuntu/debian/changelog: releasing version 20100518ubuntu608:42
cjwatsonbdmurray: do you know how to get http://reports.qa.ubuntu.com/reports/kernel-bugs/reports/rls-p-tracking-bugs.html to come up to date?08:55
gemacjwatson: what do you want us to do with uefi grub re bug 934614?09:26
ubot2Launchpad bug 934614 in grub2 "reinstall of precise breaks grub with invalid arch independent ELF magic" [High,Incomplete] https://launchpad.net/bugs/93461409:26
gemacjwatson: john johansen is also willing to help, but I am not sure what you are after from your comment?09:26
cjwatsongema: in my comment I indicated that I could use the output of boot_info_script09:27
cjwatsonwas that unclear?09:27
gemacjwatson: yes, let me read it again09:27
cjwatson"The output of 'sudo boot_info_script' from http://bootinfoscript.sourceforge.net/ (the boot-info-script package is only version 0.60 which doesn't support UEFI, so please make sure to use at least version 0.61) may be useful in tracking this down."09:27
cjwatsonand my install was UEFI, I do know the difference ;-)09:28
gemacjwatson: does your grub install and work with all the uefi bits and pieces?09:29
cjwatsonyes.09:29
gema:D09:29
gemaahh, ok09:29
cjwatsonI know the difference. :-)09:29
gemayeah, I don't so I am just making sure, bear with me09:29
cjwatsonI mean it's a bit busted, but it doesn't have the sort of fundamental breakage you're reporting09:29
gemacjwatson: ack09:29
cjwatsonwhich to me smells very strongly of a UEFI grub core image trying to load BIOS grub modules or vice versa - which is a misconfiguration09:30
cjwatsonit could for example happen if there's a prior installation and the system is booting from the wrong one, but now /boot/grub/ contains a different type of modules09:30
cjwatsone.g. you booted the installer using UEFI but are attempting to boot the installed system using BIOS - this is typically a firmware menu choice09:31
cjwatsonyou have to be consistent and boot the installer and the installed system using the same one09:31
cjwatsonof BIOS vs. UEFI09:31
gemacjwatson: how do we know which booting method the liveCD used?09:31
cjwatsonit installed grub-efi therefore it was booted using UEFI09:32
gemaok09:32
cjwatsonit is possible for it to be booted either way09:32
cjwatsonit's usually a boot menu choice09:32
gemaok, understood, we'll run the script and update results as soon as I can get that machine to reboot :)09:32
cjwatsonI expect the script will indicate that there are bits of the previous BIOS GRUB installation lying around09:32
gemaok09:33
cjwatsonwhich isn't actually an error as such, it will just mean you need to figure out how to boot the installed system using UEFI09:33
cjwatsonBTW if you considered /var/log/installer/partman sensitive, then the output of boot_info_script is too09:33
gemaok09:33
gemacjwatson: what information exactly are you looking for, I am finding difficult to convince the owner of the machine to run that script09:41
cjwatsongema: I need to know whether there are bits of a BIOS GRUB lying around somewhere, to confirm this hypothesis09:42
cjwatsongema: otherwise, the owner of the machine needs to be very clear whether they're booting the installed system using BIOS or UEFI, and (if my hypothesis is correct) learn how to boot it using UEFI09:43
gemacjwatson: his machine always tries to boot UEFI and falls back to bios if UEFI doesn't work09:50
gemacjwatson: there is no configuration setting to do anything else09:50
cjwatsonI'm afraid, then, I don't know.  It works for me.09:55
gemacjwatson: yeah, but we are not shipping your machine, so let me try contacting jjohansen and see if he can run the script10:15
cjwatsongema: I understand that we are not shipping my machine, but the error reported here is unlikely to be hardware-specific10:16
gemacjwatson: you are probably right, that is exactly why I'd like to get it fixed, let me talk to john and see if he can run the script and get us that data10:17
cjwatsongema: The error indicates that GRUB is trying to load a module which is fundamentally incompatible with the GRUB core image being run10:17
cjwatsongema: If that were a bug in the package, it would be reproducible on every UEFI install10:17
cjwatsongema: Contrariwise, since it is not reproducible on every UEFI install, that suggests that it is likely a configuration problem of some kind10:17
gemacjwatson: configuration of what?10:18
cjwatsongema: And, since the report concerns a situation of a UEFI install being performed over the top of a BIOS install, the overwhelming probability is that some bit of the prior BIOS install is being loaded10:18
cjwatsonFrom the point of view of GRUB, that's a configuration problem10:18
cjwatsonAnd if the user is unwilling to provide all the information I need, I'm afraid I can't proceed10:19
gemacjwatson: I understand that, I am trying to find a more willing user, will ping you when we have that data10:21
stgrabercjwatson: oh, "debconf-show grub-pc" indeed showed me what's wrong ;)13:12
stgrabercjwatson: I installed that machine from kvm, then rebooted on metal which is how I usually install systems remotely in that datacenter13:12
stgrabercjwatson: but grub stored the disk/by-id path to the disks, which no longer match13:12
stgrabergrub-pc/install_devices: /dev/disk/by-id/ata-QEMU_HARDDISK_QM00001, /dev/disk/by-id/ata-QEMU_HARDDISK_QM0000213:12
stgraberthese should instead be /dev/disk/by-id/ata-ST3000DM001-9YN166_S1F013HW /dev/disk/by-id/ata-ST3000DM001-9YN166_S1F01Q0B13:13
cjwatsonright, you need 'dpkg-reconfigure grub-pc' after changing disks13:13
cjwatsongrub doesn't have a safe way to tell what to do otherwise13:14
stgrabercjwatson: done, thanks. I guess I'll add a check for that in my post-install sanity check script (ensuring that everything listed in install_devices actually exists)13:21
stgraberjibel: ping14:02
stgraberjibel: could you confirm the fix for bug 772470 with your setup?14:03
ubot2Launchpad bug 772470 in ubiquity "os-prober doesn't detect Windows partition but the recovery partition instead" [High,Fix released] https://launchpad.net/bugs/77247014:03
jibelstgraber, ok will do.14:06
jibelstgraber, no difference. no wubi option and no m-a, same output for os-prober. Are there any specific conditions the system should satisfy to display these options ?16:24
stgraberjibel: argh...16:25
stgraberjibel: it needs to have ntfs partitions containing user directories16:25
stgraberjibel: can you paste /var/log/installer/debug and /var/log/syslog somewhere?16:25
stgraberjibel: during testing yesterday I had a few times where ntfs wouldn't mount causing what you're describing here16:26
jibelstgraber, aha, you're right: "Failed to mount '/dev/sda3' Operation not permitted"16:29
bdmurraystgraber: you might look at the merge proposal for bug 97379416:29
ubot2Launchpad bug 973794 in casper "Malformed 'search' entry in /etc/resolv.conf on network install" [High,New] https://launchpad.net/bugs/97379416:29
stgraberyay! not my fault ;)16:29
jibelstgraber, fixing windows first and will install again16:29
stgraberjibel: thanks16:30
stgraberjibel: apparently Windows 8 likes to suspend-to-disk on shutdown, making mounting the partition almost always impossible, so we'll need to figure something out for 12.1016:30
stgraberjibel: one option is to cancel the suspend to disk (there's an ntfs-3g option for that) but that may end up with data loss for the user16:30
stgraberbdmurray: I'll have a look, thanks16:31
jibelstgraber, yep, my test systems sometimes stops in weird states. chkdsk usually fix that.16:32
stgraberbdmurray: right, valid bug report but I don't really like the fix, changing the logic completely has a high potential of regression. I'll have a look at just fixing the sed regexp16:51
jibelstgraber, m-a works16:56
stgraberjibel: cool, did wubi work too (when the partition table is full)?16:57
jibelstgraber, no, I must change the partitioning in windows first.16:57
stgraberk16:57
stgrabergema: Can I get ssh access to a machine affected by bug 966294?17:23
ubot2Launchpad bug 966294 in ubiquity "Ubiquity loops forever from ubiquity_webcam_play" [High,Confirmed] https://launchpad.net/bugs/96629417:23
gemastgraber: I guess I can make it happen, yes, gimme a sec17:25
stgrabergema: thanks17:25
=== cyphermox_ is now known as cyphermox
CIA-32ubiquity: cjwatson * r5359 trunk/ (151 files in 3 dirs): Update translations from Launchpad.22:15
CIA-32ubiquity: cjwatson * r5360 trunk/debian/changelog: releasing version 2.10.822:57

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!