[00:20] <exlt> is the ubuntu installer synced from debian-installer regularly?  I'm just not clear about the dev path - if this was a bug introduced by d-i, or u-i, and where I might try to be most helpful
[00:22] <exlt> I don't need to pass any $LANG boot args to squeeze d-i, for instance
[00:23] <exlt> preseed works fine there
[00:44] <exlt> looking at the installer build version 20081029ubuntu80 - does this indicate a d-i fork from 10.29.2008?
[00:46]  * exlt is reading http://wiki.ubuntu.com/Installer/Development..
[01:05] <exlt> cjwatson: 20081029ubuntu79 does not show this particular behavior, if that helps at all
[01:10] <atrus> maybe the wrong place to ask, but I'm trying to follow https://help.ubuntu.com/community/Installation/FromUSBStick , but the freshly formatted drive always reports "not enough free space" when I try to write to the usb device. What can I do?
[01:12] <exlt> atrus: what size usb drive?
[01:12] <atrus> 4gig
[01:12] <exlt> hrm..
[01:13] <atrus> i format the thing in usb-creator-gtk, and it immediately says there's zero free space on the thing.
[01:14] <exlt> I don't know much about gui tools
[01:15] <atrus> is there a command-line way to get an ISO onto a usb drive? i know you can dd a .img file...
[01:15] <exlt> I would do it from cli, but don't know if that messes up the gui widget
[01:15] <atrus> i'd gladly do it all from the command line if i could figure out how.
[01:17] <atrus> exlt: how would *you* do it from the cli?
[01:18] <exlt> not sure what it's expecting to have as an end result, but I've done this with debian using syslinux and a netinst iso - looking for the docs foryou
[01:18] <atrus> i really just need a bootable usb rescue img.
[01:18] <atrus> grub's broken on my other machine, and i need a way to get in and fix it.
[01:20] <exlt> yeah, this looks about the same process - http://www.ubuntugeek.com/how-to-install-ubuntu-linux-from-usb-stick.html
[01:21] <exlt> hmm - he doesn't say anything about the syslinux config
[01:21] <exlt> lemme see if I have a usb drive with it on there and take a peek
[01:22] <exlt> oh, he's copying the isolinux dir from the iso image, so that should work
[01:24] <atrus> exlt: so follow that link? it didn't seem to say anything about how to format the usb drive's filesystem...
[01:24] <exlt> fdisk /dev/<wherever_your_usb_drive_is>
[01:25] <exlt> fdisk can be a bit daunting..
[01:25] <exlt> is it already formatted? or is that what was dorked with the gui thing?
[01:25] <atrus> it's been corrupted.
[01:26] <exlt> are there any partitions listed?
[01:26] <atrus> i think i can get through using fdisk though, i just need to know how to format it. vfat, fat32, ext2/3?
[01:26] <exlt> p - will print the partitions
[01:26] <exlt> ah, ok
[01:26] <atrus> yeah, no partitions, it recreated the partition table.
[01:27] <exlt> mkfs.msdos /dev/something1
[01:27] <exlt> uhh, maybe mkfs.vfat
[01:27] <exlt> yeah, that should probably be vfat
[01:28] <atrus> with you so far
[01:30] <atrus> copying everything from the mounted .iso into the usb...
[01:32] <exlt> cjwatson: yep - 20081029ubuntu79 works for me perfectly, however, the current build is kind of a showstopper for pxe booting - did you create a bug that I can follow, or which package would this be against and I can create one
[01:33] <exlt> atrus: not sure the steps said to copy the contents
[01:33] <atrus> er, okay, this is weird. they talk about copying over files in /install, and then running syslinux, and then copying over the iso's files again, overwriting them?
[01:34] <atrus> i'm pretty sure i'm misunderstanding step 1.
[01:34] <exlt> install dir with vmlinux and initrd.gz inside
[01:35] <exlt> there's a little image at the bottom of the entire contents of the root of the usb drive  :)
[01:35] <atrus> right, and then /install gets more stuff from the .iso image afterwards?
[01:36] <exlt> oh - did you follow the links to grab the hd-media kernel and initrd?
[01:36] <exlt> those links are *way* old
[01:36] <exlt> so will need to be adjusted for whatever version you want
[01:36] <exlt> what version are you wanting to install?
[01:37] <atrus> right, trying for karmic. i noticed the out-of-date link now :)
[01:37] <exlt> replace hardy with karmic and they should work a little better
[01:37] <atrus> i originally just assumed the /install directory from the cd would include the neccesary components.
[01:38] <exlt> nah, this is really a bootstrapping to load up the iso image from the usb drive
[01:40] <atrus> ahhh, ok ok.
[01:41] <atrus> i think that's the one piece of information i critically missed when reading this :)
[01:42] <exlt> yeah - it's a kernel and initrd with enough magic to go look for an iso file, mount it on loopback, and install from it like it's a normal CD
[01:43] <atrus> exlt: how does it know what .iso file to load?
[01:44] <atrus> ie, i'm plugging ubuntu-9.10-desktop-i386.iso in there, but it could be -server instead.
[01:45] <exlt> it will load just about any, but the kernel versions need to match, if I recall - I've done that at some point with a newer iso and it complained at me that it didn't match
[01:45] <atrus> yeah, it should match up there.
[01:47] <atrus> thanks a lot, whether this works or not when it's finished copying in. i've learned a bit more about this if nothing else.
[01:48] <exlt> that's half the fun  ;)
[01:48]  * exlt has been playing with linux since ~1997
[01:49] <atrus> ~98 for me. but syslinux and friends has always been something i've never had to dig too much into.
[01:50] <atrus> thankfully :)
[01:50] <exlt> I dropped in here with a frustrating problem that I could not figure out how to "configure" away, and it turns out we found a bug - so I help
[01:51] <exlt> never did find my usb installer - should probably make one..
[01:51] <exlt> real handy for rescues
[02:08] <exlt> cjwatson: as I thought might be the case, I don't need to pass locale or layoutcode as boot args when using auto and priority=critical - they are deferred and used from the preseed
=== __ruben__ is now known as _ruben
[09:57] <CIA-41> ubiquity: evand * r3715 ubiquity/ (4 files in 2 dirs):
[09:57] <CIA-41> ubiquity: * Move oem-config runtime debconf questions into the oem-config
[09:57] <CIA-41> ubiquity:  package.
[09:57] <CIA-41> ubiquity: * Add oem-config/early_command (run a command just before oem-config
[09:57] <CIA-41> ubiquity:  starts).
[09:59] <ev> Is the lack of oem-config/id in any templates file an oversight?
[10:00] <cjwatson> it didn't need to be there as it was only preseeded and preseeding automatically registers missing templates if it needs to.  It wouldn't hurt to add it for documentation though
[10:12] <ev> ah, I figured as much; just wanted to be sure
[10:12] <cjwatson> ev: did you see my highlight last night about console-setup?
[10:12] <ev> cjwatson: no, I'm still getting the hang of ZNC, or rather trying to get it to actually replay everything.
[10:13]  * ev digs through irclogs.ubuntu.com
[10:13] <ev> ask_debconf> will do
[10:28] <superm1> ev, if you want to glance over https://code.edge.launchpad.net/~superm1/ubiquity/mythbuntu-plugin-enhance again, i've got it to the point now that mythbuntu_ui is now entirely dead.  i'm gonna do some more testing tomorrow to make sure I didn't break gtk_ui (w/o the myth plugins loaded)in the process
[10:28] <ev> superm1: cool, I'll endeavor to fit it in today
[10:29]  * ev continues to fight against the en_GB dictionary that got installed when he reformatted
[11:01] <ara> ev, I got an error with DHCP (OEM mode) while testing hardy point release
[11:01] <ara> ev, bug 512740
[11:01] <ubottu> Launchpad bug 512740 in debian-installer "DHCP failed several times and istaller looped in the network configuration" [Undecided,New] https://launchpad.net/bugs/512740
[11:11] <ev> ara: looking now
[11:12] <ara> ev, thanks :)
[12:11] <CIA-41> ubiquity: evand * r3716 ubiquity/ (3 files in 2 dirs):
[12:11] <CIA-41> ubiquity: Support installing additional packages at runtime in oem-config, as
[12:11] <CIA-41> ubiquity: specified by oem-config/extra_packages.
[13:18] <CIA-41> ubiquity: evand * r3717 ubiquity/scripts/install.py: Fix missing parameter.
[14:30] <dmarkey> might be the wrong place to ask, but why is openoffice on a server default install
[14:32] <soren> dmarkey: It's not?
[14:33] <dmarkey> it is for me, 9.10
[14:45] <exlt> dmarkey: I found the same thing and dug around the cdrom preseeds and found:
[14:45] <exlt> d-i pkgsel/install-language-support boolean false
[14:46] <soren> dmarkey: What exact packages are being installed that you were not expecting?
[14:46] <exlt> can't remember the exact package, but there was a circular dependency that pulled in a couple openoffice packages
[14:47] <soren> "a couple of openoffice packages" != "openoffice". :)
[14:47] <exlt> well, sure, but any openoffice packages on a server is kind of useless..
[14:55] <dmarkey> exlt: ok so that must cut out a few 100 megs?
[14:56] <dmarkey> soren: openoffice, and java.
[14:56] <soren> dmarkey: There's no package called "openoffice" nor "java".
[14:57] <soren> It's much easier to work out what went wrong if I can get exact data on what the specific problem is.
[14:57] <exlt> I think it was openoffice.org-thesaurus-en-au that you are seeing, and another - they get pulled in by language-support-en, language-support-writing-en.. - that d-i line above kept them from being installed
[14:58] <soren> exlt: /Kept/ them from being installed? I thought you guys were saying this stuff /did/ get installed?
[14:58] <exlt> soren: they did with a simple preseed of:
[14:58] <exlt> tasksel tasksel/first multiselect standard
[14:59] <exlt> (which does not *exclude* install-language-support specifically
[15:00] <exlt> the CD seed files specifically exclude installation of install-language-support
[15:01] <dmarkey> but not the net install?
[15:01] <exlt> correct
[15:02] <exlt> well, I could be wrong there, perhaps - as soon as one specifies a custom preseed, that overrides any default
[15:02] <exlt> so I found that line, added it to my preseed, and no more openoffice packages
[15:02] <cjwatson> this is a bug actually, as of 9.10 this is meant to be handled automatically
[15:03] <cjwatson> I think the bug is that language-selector-common needs to be added to the server-ship seed
[15:03] <cjwatson> without it, we're falling back to old behaviour which was too stupid to work this out for itself
[15:03] <exlt> not sure about 9.10 - I'm working on 10.4
[15:04] <cjwatson> no change in that regard yet
[15:04] <cjwatson> maybe I should just put language-selector-common in standard, since the effect now will be that it's installed for everything regardless
[15:04] <cjwatson> it's tiny
[15:05] <cjwatson> yeah, I'll do that
[15:07] <cjwatson> exlt: hopefully the next daily build of lucid won't require that preseed
[15:10] <dmarkey> cjwatson: thanks
[15:11] <dmarkey> cjwatson: how did your nfsv4 experement go?
[15:11] <cjwatson> partial, I got Kerberos working but haven't quite got NFSv4 up and running yet
[15:11] <dmarkey> for insecure use, nfsv4 doesnt require kerberos
[15:11] <cjwatson> that's nice ... :-)
[15:12] <cjwatson> I want to have the option of using it off my home network
[15:12] <cjwatson> and I'd been meaning to investigate Kerberos for years anyway
[15:19] <dmarkey> yea, and LDAP too?
[15:20] <cjwatson> I'm not that masochistic
[15:20] <ev> haha
[15:21] <dmarkey> oh right, they usually go hand in hand
[15:21] <dmarkey> kerberos/nfsv4/ldap that is
[18:07] <atrus> exlt: for what it's worth, that didn't work in the end. blinking cursor. (shrug) I ended up downloading a debian usb image, since they actually distribute the .imgs without forcing you to make your own out of an iso. (which wouldn't have been a problem if usb-creator worked)
[18:41] <exlt> atrus: I did the same last night on an old 256M stick ;) - zcat boot.img.gz > /dev/sdX
=== cjwatson_ is now known as cjwatson