[05:12] <whiteheat> hi, anybody had issues installing oneiric? Works fine from live CD, then install, then reboot, then wont start X as it cant find driver for nvidia
[05:31] <whiteheat> tried modprobing nv and nouveau driver, failed. Tried compiling module, fails - suspect it is due to not being a 2.6 kernel, but a 3.0 kernel
[11:41] <CIA-2> ubiquity: cjwatson * r5157 trunk/ (4 files in 3 dirs):
[11:41] <CIA-2> ubiquity: Don't display "Ready when you are..." when the installer isn't actually
[11:41] <CIA-2> ubiquity: waiting for user actions.
[14:40] <CIA-2> ubiquity: cjwatson * r5158 trunk/ (78 files in 3 dirs):
[14:40] <CIA-2> ubiquity: When waiting for user actions, just hide the progress bar and associated
[14:40] <CIA-2> ubiquity: widgets rather than displaying "Ready when you are...", to avoid
[14:40] <CIA-2> ubiquity: confusion associated with a progress bar being displayed when the system
[14:40] <CIA-2> ubiquity: is in fact waiting for user input (LP: #670409, #925427).
[14:47] <stgraber> cjwatson: hi! I have what should hopefully be a quick and easy partitioning question for you ;) Yesterday I got access to the upgraded version of one of my datacenter servers (personal one) and it came with 2x3TB hard disks. I'm trying to install Precise on it having the system and bootloader/boot partitions in a software RAID1. My initial try with a regular ext4 partition over RAID1 (as I'd do with a non-GPT disk) made grub fail and I'
[14:48] <cjwatson> truncated at "made grub fail and I'"
[14:48] <stgraber> gah
[14:48] <stgraber> made grub fail and I'm now trying to understand what I need to change to make it work ;)
[14:49] <cjwatson> Hm.  Sounds like we have a grub bug to fix then ...
[14:49] <cjwatson> You can often debug this kind of thing in userspace with grub-probe
[14:50] <stgraber> so what I currently have on both disks is a 1GB RAID1 partition and a 1.5TB RAID1 partition. The first one (md0) contains an ext4 filesystem used for /boot and the second one (md1) is an LVM PV used for the system. Is that supposed to work on GPT or am I missing something?
[14:52] <cjwatson> That should in principle all work on GPT, yes.
[14:53] <cjwatson> Maybe we have a lurking 32-bit variable somewhere or something.
[14:53] <cjwatson> In fact if you only have 1.5TB in use then it should *definitely* work.
[14:54] <stgraber> ok, will do another install using that and then grab whatever you need to debug (the disks are blank, so I'm currently limited to a rescue system running kvm on the disks)
[14:54] <cjwatson> Does GRUB fail to install or fail to boot?
[14:55] <stgraber> fail to install, complaining about embedding space IIRC
[14:55] <cjwatson> Did you create a BIOS Boot Partition?
[14:55] <cjwatson> http://www.gnu.org/software/grub/manual/grub.html#BIOS-installation
[14:56] <stgraber> ok, that's what I'm missing then. Can I create that on top of software RAID1?
[14:57] <cjwatson> I think it needs to be directly on GPT
[14:57] <cjwatson> In your situation I'd probably create one on each disk
[14:58] <stgraber> right, then run grub-install twice to make sure both would work (degraded boot)?
[15:03] <cjwatson> grub-install ought to be run to both disks, yes
[16:09] <bdmurray> ev: could you look at bug 925779?
[16:09] <ubot2`> Launchpad bug 925779 in wubi "Wubi shouldn't fail when uninstalling if the BCD entry is missing" [Undecided,New] https://launchpad.net/bugs/925779
[16:10] <ev> bdmurray: I don't have time to fix it right now, but I've triaged it
[16:10] <ev> bcbc's point is correct
[16:11] <bdmurray> ev: it sounds to me like he is personally stuck though
[16:11] <bdmurray> ev: which is too bad since he does lots of wubi bug triage
[16:14] <ev> okay, followed u
[16:14] <ev> up
[16:15] <bdmurray> ev: great, thanks
[16:16] <ev> no problem
[16:44] <bdmurray> stgraber: have you seen bug 924998?
[16:44] <ubot2`> Launchpad bug 924998 in network-manager "no dns resolution when wireless connection is enabled by ubiquity" [Undecided,New] https://launchpad.net/bugs/924998
[16:46] <stgraber> bdmurray: that'd be one of cyphermox
[16:48] <cyphermox> I'll get to it soon
[16:51] <stgraber> bdmurray: one thing that'd usually be useful on top of the current attachments would be /run/resolvconf/resolv.conf and something checking where /etc/resolv.conf points to (and more importantly, making sure it's a symlink)
[16:51] <stgraber> (in this case, I'd expect the syslog to contain enough information to figure this out, but having the extra information directly from apport would be nice)
[16:52] <bdmurray> stgraber: for ubiquity or more packages than just it?
[16:52] <stgraber> bdmurray: whenever we deal with something network related, so ubiquity, NetworkManager, resolvconf and ifupdown at least
[16:57] <bdmurray> stgraber: there is a general attach_network function in apport that could gather that and then be used by ubiquity and other pacakges
[16:57] <stgraber> that sounds good
[18:05] <bdmurray> cjwatson: I was looking at bug 912031 and noticed that partman-auto information isn't included in ubiquity bug reports.  I could add it in the apport package hook but do you think needing that info comes up often?
[18:05] <ubot2`> Launchpad bug 912031 in partman-auto "ubiquity crashed with TypeError in partman_column_name(): argument of type 'NoneType' is not iterable" [High,Fix released] https://launchpad.net/bugs/912031
[18:10] <cjwatson> bdmurray: /var/log/partman is included
[18:11] <cjwatson> bdmurray: I can't think what other information you might be talking about
[18:12] <bdmurray> cjwatson: package version information sorry
[18:12] <bdmurray> cjwatson: it doesn't appear in Dependencies.txt
[18:13] <cjwatson> bdmurray: Not needed, it's part of the ubiquity binary package, not a separate .deb
[18:13] <cjwatson> partman-auto is only a udeb so ubiquity can't depend on it in the usual way
[18:13] <cjwatson> see the "Automatic update of included source packages" stuff in ubiquity's changelog
[18:14] <cjwatson> But we fix bugs in the included source packages rather than forking them for ubiquity
[18:14] <cjwatson> It's a slightly odd arrangement but it works :)
[18:15] <bdmurray> sure but then "This bug was fixed in the package partman-auto - 93ubuntu19" isn't very helpful
[18:15] <cjwatson> That's OK
[18:15] <cjwatson> Because, I mean, it *was*
[18:15] <cjwatson> it's not IME worth the effort of adding ubiquity tasks, and I generally leave an explicit comment in the odd case where users might get confused
[18:16] <cjwatson> generally speaking it doesn't seem to have been a problem
[18:18] <bdmurray> I guess I could write a bug pattern for partman_column_name and ubiquity version < 2.9.16
[18:25] <cjwatson> bdmurray: ubiquity versions and the versions of its included sources map one-to-one, so that kind of transformation is safe, yes
[18:26] <bdmurray> I think the challenge is getting from reading bug 9120931 seeing the changelog entry and *knowing* how ubiquity is built to find out which version of ubiquity it was fixed in
[18:28] <cjwatson> I'd be happy to comment explicitly on hot bugs to that effect, but I don't think we need it in general
[18:28] <cjwatson> unless you're talking about doing it automatically, in which case we could build some kind of lookup table from ubiquity's source history
[18:29] <bdmurray> I'll think about it more
[18:51] <bdmurray> stgraber: I have a system experiencing bug 645449 if there is something useful I could do / gather
[18:51] <ubot2`> Launchpad bug 645449 in ubiquity "Ubiquity hangs at Keyboard layout if you use keyboard to navigate / select" [High,Confirmed] https://launchpad.net/bugs/645449
[19:04] <stgraber> bdmurray: do you happen to have a reproducable way of triggering it?
[19:05] <bdmurray> stgraber: Using the keyboard to move up and down in the layout choice multiple times seems to do it for me
[19:05] <bdmurray> stgraber: I've recreated it more than once
[19:06] <stgraber> bdmurray: good, that was the biggest issue with that bug, being able to actually reproduce it somewhat reliably ;)
[19:06] <stgraber> bdmurray: can you try it with ubiquity in debug mode?
[19:06] <bdmurray> also bug 913779 has a good description
[19:06] <ubot2`> Launchpad bug 913779 in ubiquity "Fail to choose keyboard type when installing Precise on Macbook Air 4,1 (dup-of: 645449)" [Undecided,New] https://launchpad.net/bugs/913779
[19:06] <ubot2`> Launchpad bug 645449 in ubiquity "Ubiquity hangs at Keyboard layout if you use keyboard to navigate / select" [High,Confirmed] https://launchpad.net/bugs/645449
[19:09] <stgraber> bdmurray: indeed, that's a pretty good description. Other than looking for a python stack trace in the usual logs, I can't think about something else that'd be useful at this point, but it might be worth actually targeting the bug for beta-1
[19:10] <stgraber> bdmurray: my plan is to switch to 90% installer work (bug fixing) post feature freeze (with the remaining 10% being networking bugs), so I'll be sure to look at this one (especially as Danial Stoni has been poking me quite a few times, on LP, IRC, direct e-mail and through the Ubuntu CH mailing list ...)
[19:21] <bdmurray> stgraber: this is all I see in debug
[19:21] <bdmurray> /usr/lib/ubiquity/ubiquity/misc.py:572: UserWarning: set_indicator_keymaps: this function currently does not work
[19:28] <stgraber> right, this one is expected (we explicitly turned it off in the code) and shouldn't be the source of the problem (because we changed it in Oneiric and the bug was already there long before)
[19:29] <bdmurray> stgraber: I've added a screenshot of syslog
[19:56] <bdmurray> cjwatson: there are actually a couple of bugs with 2.9.16 still crashing in ubi-partman
[19:56] <bdmurray> bug 926248 and bug 926150
[19:56] <ubot2`> Launchpad bug 926248 in ubiquity "daily current 3rd Feb. doesn't allow 'select other method' on setup menu" [Undecided,New] https://launchpad.net/bugs/926248
[19:56] <ubot2`> Launchpad bug 926150 in ubiquity "Ubiquity crush on disk partitioning ( unable to load list )" [Undecided,New] https://launchpad.net/bugs/926150
[20:14] <cjwatson> bdmurray: I'd be surprised if there weren't - that's a large amount of code to crash in
[20:14] <cjwatson> I'll look at those at some point but it's Friday evening :)
[20:16] <cjwatson> ok, so I guess I didn't entirely manage to figure out why on earth None values are sneaking in there
[20:18] <cjwatson> (I realise now you're talking about the exact same symptom rather than just any crash in ubi-partman, which is what I thought you meant at first ...
[20:18] <cjwatson> )
[21:24] <CIA-2> ubiquity: cjwatson * r5159 trunk/ (debian/changelog ubiquity/parted_server.py):
[21:24] <CIA-2> ubiquity: Mark file descriptors used for communication with parted_server
[21:24] <CIA-2> ubiquity: FD_CLOEXEC.
[21:24] <cjwatson> bdmurray: ^- wasn't necessary to fix Kate's system, but was one of the things I found as a possibility when debugging it; we'll see if that helps
[21:25] <cjwatson> basically the same kind of fix in a slightly different place
[21:25] <bdmurray> cjwatson: got it, thanks
[22:42] <bdmurray> stgraber: is the boot screen where you choose try ubuntu and keyboard and languages customizable?  bug 924901
[22:42] <ubot2`> Launchpad bug 924901 in casper "xubuntu live shows ubuntu's boot screen" [Undecided,New] https://launchpad.net/bugs/924901
[22:43] <stgraber> bdmurray: I'm not sure. Edubuntu simply doesn't have that screen. Anyway, it's most likely a gfxboot question more than a casper question (unless I don't understand which purple screen he's talking about)
[22:44] <bdmurray> I interpret keyboard + man to mean the gfxboot screen
[22:45] <stgraber> right, so that's the screen shown before the menu (unless you don't press anything, then you boot directly). So should definitely be re-assigned to gfxboot, as for the config, I really have no idea.