[01:55] <CIA-59> base-installer: cjwatson * r333 ubuntu/ (debian/changelog kernel/i386.sh): Add support for AMD family 17 on i386.
[04:05] <kirkland> cjwatson: I'm seeing some behavior in grub-installer that I didn't expect....
[04:05] <kirkland> cjwatson: syslog of a RAID install: http://pastebin.ubuntu.com/47315/
[04:06] <kirkland> cjwatson: the log-blob starting at line 8444 shows that grub-installer finds and executes /sbin/grub-install in the chrooted target OS
[04:07] <kirkland> cjwatson: I think this actually makes my RAID and md-related changes to grub-installer moot
[04:07] <kirkland> cjwatson: however, grub-installer is still calling /sbin/grub-install with "(hd0)" as the argument, so grub-install is not actually writing the multiple MBR's
[04:10] <kirkland> cjwatson: actually, this might be a problem too...
[04:10] <kirkland> Sep 15 22:27:42 grub-installer: info: Identified partition label for /dev/md0: loop
[04:10] <kirkland> "loop" is not the correct label
[04:42] <kirkland> cjwatson: okay, i think i may see the heart of the problem...
[04:42] <kirkland> if type dmraid >/dev/null 2>&1; then
[04:42] <kirkland> ...
[04:43] <kirkland> elif type multipath >/dev/null 2>&1; then
[04:43] <kirkland> ...
[04:43] <kirkland> elif type mdadm >/dev/null 2>&1; then
[04:43] <kirkland> ...
[04:43] <kirkland> fi
[04:43] <kirkland> cjwatson: those are not mutually exclusive
[04:43] <kirkland> cjwatson: my install has dmraid such that "type" returns 0, but i really want the mdadm block to execute
[05:08] <TheMuso> kirkland: What package?
[05:08] <kirkland> TheMuso: grub-installer
[05:08] <kirkland> TheMuso: I'm testing a patch now....
[05:09] <TheMuso> kirkland: Ah ok. I have yet to debug why it doesn't do its dmraid thing properly, so if you can solve that for yourself, and assuming it gets uploaded, I will base my work from what you've done.
[05:09] <kirkland> TheMuso: sounds good, i'll pass my patch by you here, if it works
[05:09] <TheMuso> kirkland: Ok.
[05:11] <TheMuso> kirkland: I see what you mean about faulty logic, especially since dmraid is aprt of the disks now.
[05:11] <kirkland> TheMuso: oooh, did that change recently?
[05:11] <kirkland> TheMuso: I swear this was working for me earlier in Intrepid.......
[05:11] <TheMuso> kirkland: Yes, when I seeded dmraid.
[05:11] <TheMuso> As part of my dmraid inclusion work.
[05:15] <kirkland> TheMuso: \o/ it works :-)
[05:15] <TheMuso> kirkland: Nice.
[05:18] <kirkland> TheMuso: patch attached to https://bugs.launchpad.net/ubuntu/+source/grub-installer/+bug/270758
[05:18] <kirkland> TheMuso: I'm making a bzr branch now
[05:41] <TheMuso> kirkland: Ok.
[05:41] <evand> Is anyone else able to confirm that persistence support in casper is broken at the moment
[05:42] <TheMuso> evand: Whats needed to test? I have a USB key I can dig up, but have never played with it before.
[05:44] <TheMuso> kirkland: Glancing at the diff, I like the extra bit of logic you've added there. My only concern is that a user *MAY* wish to use a dmraid device, multipath, and mdadm simultaneously, but I highly doubt it. :p
[05:45] <evand> TheMuso: you can either make an ext2 partition on that disk with the label casper-rw, or you can put the full CD on the USB disk and add a casper-rw loopback file in the root.  In either case, add 'persistent' to the kernel command line.
[05:45] <evand> https://wiki.ubuntu.com/LiveUsbPendrivePersistent
[05:45] <TheMuso> evand: Ok, I'll have to do the former, since my key is only 128MB.
[05:46] <TheMuso> I'll get back to you in a bit with the results.
[05:46] <evand> ok, much appreciated
[05:53] <TheMuso> ...or maybe not. Looks like the key is dead.
[05:53] <evand> ouch
[05:53] <evand> thanks the same though
[05:53] <TheMuso> np
[05:54] <TheMuso> It wasn't a critical key. I only ever used it for testing/experiments.
[05:54] <TheMuso> I'd rather carry everything around on either an external HD or on my notebook. Those keys are too easy to lose.
[05:57]  * evand puts everything in the cloud, lest he lose important information to a disk failure.
[06:02] <persia> evand: Everything?  Even your secret keys?
[06:03] <evand> ok, so not those :)
[06:05] <persia> evand: I understand.  I recently travelled with a formatted notebook, and an install CD.  Was functional again as soon as I had a network connection.
[06:06] <evand> nice
[06:11] <TheMuso> Its alright if sed network connection is fast.
[06:12] <TheMuso> I don't think cloud computing will be worthwhile in Australia till we get some deacent speeds.
[06:12] <evand> heh
[06:12] <evand> fair enough
[06:12] <TheMuso> And local servers that are part of the cloud.
[06:12] <persia> TheMuso: You just need a local cloud.
[06:19]  * evand kicks CIA-59 
[06:19] <CIA-59> ow
[06:19] <evand> sure, respond to that and not my attempt to post a commit to you.
[06:20] <TheMuso> haha
[08:13] <saispo> hi
[09:42] <saispo> cjwatson: ping ?
[09:57] <cjwatson> saispo: yes?
[09:58] <saispo> cjwatson: can you give me a little help on an lvm partition scheme ?
[09:59] <saispo> it didn't work and don't understand and see no mistake :/
[10:01] <saispo> you can see it at : http://pastebin.com/m79de50ac
[10:01] <saispo> thks
[10:05] <cjwatson> saispo: you made lots of that up, didn't you? :-) I don't think device{ /dev/sda } can possibly be documented anywhere.
[10:05] <saispo> i see this in the partman recipe :)
[10:06] <cjwatson> I don't think I've ever heard of in_vg{ } or lv_name{ } either
[10:06] <cjwatson> which partman recipe?
[10:06] <saispo> http://d-i.alioth.debian.org/svn/debian-installer/installer/doc/devel/partman-auto-recipe.txt
[10:07] <cjwatson> saispo: ah. I think that's from a newer version of partman-lvm and partman-auto-lvm than Ubuntu has.
[10:07] <cjwatson> please don't follow Debian documentation on Ubuntu, you'll just get yourself confused
[10:07] <saispo> ok
[10:07] <cjwatson> it looks neat, but we don't have that stuff yet, sorry
[10:08] <saispo> ok
[10:08] <saispo> no lvm partition scheme work on Ubuntu ?
[10:08] <cjwatson> sure, just not multiple-disk ones
[10:08] <saispo> i have only one disc
[10:08] <saispo> but needed two lv in a vg
[10:08] <cjwatson> yeah, I know, but that's the multiple-disk syntax
[10:08] <cjwatson> let me finish please
[10:08] <saispo> ok
[10:09] <cjwatson> you write recipes as normal (i.e. largely as if no LVM were involved) - see http://bazaar.launchpad.net/%7Eubuntu-core-dev/debian-installer/ubuntu/annotate/963?file_id=21442%4048c42b26-1dd6-0310-b98f-a58d8bce7237%3Atrunk%252Finstaller%3Adoc%252Fdevel%252Fpartman-auto-recipe.txt
[10:10] <cjwatson> tag the partitions in the recipe that are allowed to be on LVM with $lvmok{ }
[10:10] <cjwatson> that's it
[10:10] <cjwatson> there are some specifics around line 88 of the link I just posted
[10:11] <saispo> yes, i see :)
[10:11] <cjwatson> but $defaultignore{ } and $lvmignore{ } are only important if you want to write one recipe that'll work for both LVM and non-LVM, which you probably don't
[10:11] <cjwatson> by default, it will put everything in the one VG for you, so you don't need to use the more complicated newer syntax
[10:12] <saispo> ok
[10:13] <cjwatson> looks like that new syntax will be useful for kickstart; thanks for mentioning it
[10:13] <cjwatson> sorry I accused you of making it up :)
[10:14] <saispo> no prob :)
[10:15] <saispo> will write a new partition scheme without lvm, i don't need it anymore for my customer request :)
[10:15] <saispo> thanks to take some times for me :)
[10:34] <cjwatson> kirkland: thanks, uploading
[10:54] <cjwatson> evand: is bug 182004 in fact in progress (apparently since January ...) or would you mind if I took it over at some point? I have some ideas
[10:54] <cjwatson> evand: it looks a bit more complex than just partman-auto/init_automatically_partition
[14:01] <evand> please go ahead
[14:04] <foka> cjwatson, Hello!  While edubuntu is usually built as an addon CD nowadays, I am wondering if debian-cd/data/{hardy,intrepid}/preseed/edubuntu/edubuntu.seed needs to be modified?
[14:04] <foka> cjwatson, edubuntu-desktop got renamed to edubuntu-desktop-addon?
[14:05] <foka> cjwatson, I mean the "tasksel tasksel/first   multiselect edubuntu-desktop, edubuntu-server" line.
[14:05] <foka> cjwatson, Thanks!
[14:07] <cjwatson> foka: We don't use it any more ourselves, but indeed so. Done, thanks.
[14:09] <foka> cjwatson, Wow, that was quick.  :-)  Thanks!  That was one problem I ran into: edubuntu-desktop stuff didn't get installed (but I wasn't sure if that was the change that did the trick), and it seems tasksel just quietly skipped it and left no warning in /var/log/syslog.
[14:10] <cjwatson> yeah, I don't think it warns about unknown tasks unfortunately
[14:11] <foka> cjwatson, I would like to look into it deeper some time; should I look into the "tasksel" package, or into a certain d-i module, or is it not so easily fixed?  :-)
[14:12] <cjwatson> tasksel
[14:12] <cjwatson> it's a bit messy because at one point tasksel.pl itself wasn't technically able to use debconf; now I think it could, but nobody's rewritten the workarounds
[14:13] <cjwatson> so there's a bizarre split between tasksel.pl and tasksel-debconf
[14:13] <foka> cjwatson, I see.  Thanks for the great hint!
[16:01]  * cjwatson is rather liking the shape of the debian-installer and ubiquity graphs on http://people.ubuntu.com/~bryce/Plots/
[16:01] <cjwatson> both of them are at their lowest levels all year
[16:02] <evand> very nice
[16:04] <cjwatson> evand: aiming for a usb-creator upload before slangasek wakes up and freezes for alpha-6? :-)
[16:05] <cjwatson> hmm, you haven't pushed your changes I think
[16:05] <evand> for usb-creator?  It's a bound branch.
[16:06] <cjwatson> ... but mine isn't and I was using 'bzr up'. Whoops.
[16:06] <evand> re upload> I'm going to try to squeeze one more upload in before he calls freeze, assuming he does that later than an hour from now.
[16:07] <cjwatson> I suspect you can get away with it
[16:08] <evand> it should be fairly stable and functional at the moment, though I need to investigate why persistence support in casper isn't working.
[16:14] <cjwatson> UI looks loads better
[16:14] <cjwatson> does the Format button do anything? there's no feedback
[16:16] <cjwatson> oh, I suppose it only works if I select the partition rather than the disk
[16:18] <evand> indeed, I need to put in a progress dialog for that
[16:19] <cjwatson> I think the problem is that it offered me the disk at all
[16:19] <evand> it should only offer the disk if there are no partitions on it
[16:20] <cjwatson> that wasn't true in this case, there was already a partition
[16:20] <cjwatson> if it's helpful, I inserted the stick after starting usb-creator
[16:21] <evand> hrm, what's your partition table look like?  Perhaps I can reproduce this.
[16:21] <evand> after starting usb-creator> slightly different code path, but it shouldn't make too much of a difference.
[16:22] <cjwatson> oof, it seems to have finished installing and still be running but have no windows up
[16:22] <evand> it's hidden
[16:23] <evand> you most likely clicked on something else
[16:23] <cjwatson> definitely not
[16:23] <evand> oh?  yikes.
[16:23] <evand> I haven't seen that bug in a good long while in my tests.
[16:23] <cjwatson> I've minimised everything; there are windows shown by alt-tab but actually alt-tabbing to them doesn't cause them to be displayed
[16:23] <cjwatson> last few things in the output:
[16:23] <cjwatson> [16:20:27] device_udi: /org/freedesktop/Hal/devices/volume_uuid_3433_3231
[16:23] <cjwatson> [16:20:27] num_changes: 2
[16:23] <cjwatson> [16:20:27] change: volume.mount_point
[16:23] <cjwatson> [16:20:27] change: volume.is_mounted
[16:24] <evand> is there a traceback in /root/.usb-creator.log (I know, I need to fix the write location)?
[16:24] <cjwatson> no
[16:25] <cjwatson> http://paste.ubuntu.com/47481/ from two successive invocations (I thought the first one had crashed, but actually this had happened to it)
[16:25] <cjwatson> http://paste.ubuntu.com/47482/ <- partition table
[16:27] <evand> odd, I have quite a similar setup with a 4G disk.  I'll see if I can get to the bottom of this quickly.
[16:28] <cjwatson> thanks
[18:24] <Goosemoose> any ideas on why the following brings me to the terminal in preseed.cfg instead of installing edubuntu-desktop? tasksel tasksel/first multiselect standard, edubuntu-desktop
[18:29] <persia> I suspect you want to change to edubuntu-desktop-addon, although I may be mistaken
[18:32] <Goosemoose> hmm, what i list the possibilites edubuntu-desktop is one
[18:32] <Goosemoose> nothing about addon
[18:32] <Goosemoose> maybe i need to install it as a package as well you mean?
[18:35] <superm1> i'm pretty sure that is expecting a task name, and the task is indeed called edubuntu-desktop-addon
[18:35] <superm1> just take a look at tasksel --l | grep edubuntu
[18:36] <Goosemoose> u edubuntu-server       Edubuntu server
[18:36] <Goosemoose> u edubuntu-desktop-kde  Edubuntu KDE desktop
[18:36] <Goosemoose> u edubuntu-desktop-addon        Edubuntu desktop
[18:36] <Goosemoose> damn, that's changed since feisty! no wonder, i was looking at the old docs
[18:36] <Goosemoose> ok so what's the difference between kde and addon?
[18:37] <superm1> one is gtk based and the other is kde based
[18:38] <superm1> er gnome based vs kde based is a better way to put it
[18:38] <Goosemoose> ok
[18:38] <Goosemoose> well one down one problem to go
[18:38] <Goosemoose> then im finally done with thi
[18:38] <Goosemoose> this
[18:39] <Goosemoose> do you know anything about likewise?
[18:40] <superm1> not familiar with it, sorry
[18:41] <Goosemoose> no problem
[18:52] <CIA-50> grub-installer: cjwatson * r743 ubuntu/ (debian/changelog grub-installer): merge from lp:~kirkland/grub-installer/270758
[18:57] <Goosemoose> anyone with experience with likewise? I have it joining the domain via preseed.cfg over the network. The installer logs show it's sucessfully. domainjoin-cli query shows it's joined, AD shows it's joined, but I can't log in to the domain. If I leave and rejoin it works fine. Any ideas on why that would be?
[19:00] <evand> Goosemoose: #ubuntu-server would be a better place for that question.
[19:03] <Goosemoose> evand: asked it there, no answer
[19:07] <evand> odd
[19:16] <Goosemoose> 123 people there, no one talks
[20:44] <kirkland> Goosemoose: ask dendrobates about likewise
[21:00] <Goosemoose> ok kirkland