UbuntuIRC / 2008 /11 /20 /#ubuntu-installer.txt
niansa
Initial commit
4aa5fce
=== foka_ is now known as foka
[09:44] <CIA-2> usb-creator: evand * r64 usb-creator/debian/changelog: releasing version 0.1.11
[10:11] <CIA-2> usb-creator: evand * r65 usb-creator/setup.py: Bump to 0.1.12
[12:43] <CIA-2> partman-base: cjwatson * r89 hardy-proposed/debian/changelog: releasing version 114ubuntu6
[13:00] <CIA-2> grub-installer: cjwatson * r759 ubuntu/ (debian/changelog otheros.sh):
[13:00] <CIA-2> grub-installer: Use rootnoverify rather than root when chain-loading Microsoft operating
[13:00] <CIA-2> grub-installer: systems (LP: #10661).
[13:21] <CIA-2> installation-guide: cjwatson * r436 ubuntu/ (687 files in 191 dirs): merge from Debian 20081113
[13:21] <ubottu> Error: Could not parse data returned by Debian: HTTP Error 404: No such bug (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=20081113;mbox=yes)
[13:34] <CIA-2> installation-guide: cjwatson * r437 ubuntu/debian/changelog: document kernelversion change
[15:13] <CIA-2> installation-guide: cjwatson * r438 ubuntu/ (debian/changelog en/welcome/what-is-ubuntu.xml): Bump GNOME version to 2.26.
[15:15] <CIA-2> installation-guide: cjwatson * r439 ubuntu/ (build/entities/common.ent debian/changelog): Bump release version and names for Jaunty.
[16:04] <superm1> cjwatson, a bunch of new options showed up on the mythbuntu live disk daily (Graphical install, graphical expert install, graphical rescue mode) i'm not sure if they are on ubuntu daily too, but they don't/won't work. do you know if it is an oversight or that it will only be applicable to mythbuntu disks?
[16:05] <CIA-2> partman-target: evand * r738 ubuntu/debian/changelog: releasing version 58ubuntu1
[16:39] <cjwatson> superm1: oversight during the merge, it'll get fixed, thanks for letting me know
[16:39] <cjwatson> superm1: shouldn't be showing up just yet at any rate
[17:56] <kirkland> cjwatson: hiya, i've made some good progress on encrypted home, added a "-a" option to ecryptfs-setup-private to set up all of the user's homedir to be encrypted
[17:56] <kirkland> cjwatson: this will be easy to stub into the installer
[17:56] <kirkland> cjwatson: but I'm struggling with the adduser utility
[17:57] <kirkland> cjwatson: it doesn't take the user's chosen password directly
[17:57] <kirkland> cjwatson: rather, it outsources that operation to passwd through a call to the shell
[17:57] <kirkland> cjwatson: as such, i can't exactly obtain/feed that chosen password to ecryptfs-setup-private
[17:59] <kirkland> cjwatson: the only workaround i can think of is to move some of that logic to be handled by pam_ecryptfs on first login
[17:59] <cjwatson> passwd uses pam to set the password as well
[17:59] <cjwatson> can't you intercept it at that point?
[18:00] <kirkland> cjwatson: ah, yeah
[18:00] <kirkland> cjwatson: okay, i think i can handle that
[18:00] <cjwatson> superm1: (fixed for the next build)
[18:00] <superm1> cjwatson, great thanks.
[18:00] <kirkland> cjwatson: also, regarding the installer, for this option ....
[18:01] <kirkland> cjwatson: i'd like to target replacing the "Set up an encrypted private directory" functionality with "Encrypt your entire home directory" option
[18:01] <kirkland> cjwatson: i'll keep the functionality to just encrypted a private dir, that you can setup later
[18:01] <cjwatson> kirkland: ok, that's easier to describe anyway
[18:02] <kirkland> cjwatson: even make a clicky shortcut under System->Preferences->Encryption or something
[18:02] <kirkland> cjwatson: but i think all-or-nothing will be easier from the installer POV
[18:02] <cjwatson> it'll also be easier to wedge into ubiquity's UI
[18:02] <kirkland> cjwatson: also, i think I'd like to default to randomly generated passphrase
[18:02] <kirkland> cjwatson: in the installer, anyway
[18:02] <cjwatson> fine by me, it's retrievable anyway right?
[18:02] <kirkland> cjwatson: also cuts down on the text
[18:02] <kirkland> cjwatson: yes
[18:03] <cjwatson> particularly if you make it retrievable using a UI in System->Preferences
[18:03] <kirkland> cjwatson: again, will put together a clicky util to retrieve it
[18:03] <cjwatson> would it be possible to land the UI fairly early?
[18:03] <kirkland> cjwatson: yes, assuming that either a) i can get some help from the desktop folks
[18:03] <cjwatson> I'd like for that bit not to end up getting left out due to lack of time, if we're relying on it for sanity in various places
[18:03] <cjwatson> i.e. consider in the spec what the consequences of partial implementation are
[18:04] <kirkland> cjwatson: or b) that creating a desktop link that runs the existing utilities in a popup terminal is acceptable
[18:04] <cjwatson> b) won't be, I wouldn't expect
[18:04] <kirkland> cjwatson: /me didn't figure :-P
[18:04] <cjwatson> grab desktop folks sooner rather than later
[18:04] <kirkland> cjwatson: okay, i'll try to get pitti to help, or identify someone who can
[18:05] <kirkland> cjwatson: lastly, i don't plan on providing any automated tools to "migrate" a user's homedir from not-encrypted to encrypted ....
[18:05] <cjwatson> ok, sounds good
[18:05] <cjwatson> send it by mail to pitti+Keybuk so that it gets on the UDS agenda
[18:05] <kirkland> cjwatson: i'll document how to do it from runlevel 1, as root, if anyone wants
[18:05] <cjwatson> that's fine by me; perhaps we can add language to the installation guide
[18:06] <kirkland> cjwatson: but there's too much that could go wrong with risk of a user being logged in, reading/writing data in either source or dest
[18:06] <kirkland> cjwatson: sure
[18:08] <CarlFK> jaunty, alt installer, PATA controler: Nov 20 10:54:53 disk-detect: Serial ATA RAID disk(s) detected. # Ask the user whether they want to activate dmraid devices.
[18:08] <CarlFK> report to lp?
[18:08] <cjwatson> CarlFK: --verbose
[18:09] <CarlFK> cjwatson: um... where? (that text came from /var/log/syslog
[18:09] <cjwatson> CarlFK: no, I mean "can you be more verbose"
[18:09] <CarlFK> ah.. there is no pata or raid in this box
[18:10] <CarlFK> installer is asking me if I want to activate the detected sata raid
[18:10] <CarlFK> er, no sata or raid
[18:10] <cjwatson> oh. could you run 'dmraid -c -s' from a shell please?
[18:11] <CarlFK> # dmraid -c -s
[18:11] <CarlFK> no raid disks
[18:11] <cjwatson> is that the exact output?
[18:11] <cjwatson> you haven't changed capitalisation or anything?
[18:11] <CarlFK> yup (cut/paste)
[18:11] <cjwatson> ok, yes please file a bug on hw-detect and I'll get that fixed
[18:14] <CIA-2> hw-detect: cjwatson * r93 ubuntu/ (debian/changelog disk-detect.sh): Make dmraid logging a bit neater.
[18:16] <cjwatson> CarlFK: actually, no need for a bug
[18:17] <CarlFK> cjwatson: ok
[18:18] <cjwatson> fix on its way, thanks for the early report
[18:18] <CarlFK> you're welcome
[18:18] <CIA-2> hw-detect: cjwatson * r94 ubuntu/ (debian/changelog disk-detect.sh): 'dmraid -c -s' changed its output format; cope with both old and new.
[18:23] <CIA-2> hw-detect: cjwatson * r95 ubuntu/debian/changelog: releasing version 1.70ubuntu2
[18:57] <cr3> who takes care of updating the netboot images on archive.u.c: http://archive.ubuntu.com/ubuntu/dists/intrepid/main/installer-i386/current/images/netboot/
[18:59] <cjwatson> cr3: me, but we don't update /dists/intrepid/ after release. why?
[19:00] <cr3> cjwatson: I was looking for the netboot images to test the kernel in -proposed, but I didn't find anything under /dists/intrepid-proposed
[19:00] <cjwatson> haven't uploaded one, that's why ;-)
[19:00] <cjwatson> I suppose I can do so, let me just pop a few other things off my stack
[19:01] <cr3> cjwatson: sure thing, I'll work on jaunty in the meanwhile. just drop me a ping
[19:15] <CIA-2> installation-guide: cjwatson * r440 ubuntu/ (3 files in 2 dirs): Re-enable documentation of the GTK frontend.
[19:20] <CarlFK> cjwatson: "base-installer/section/setup_dev does not exist" tail syslog: http://dpaste.com/92531/ - same install that had the raid problem, any chance it is a side effect?
[19:23] <cjwatson> CarlFK: minor bug but it's not what's causing the failure, despite being immediately before it
[19:23] <cjwatson> doesn't seem like a plausible side-effect to me
[19:24] <CarlFK> didn't think so, just making sure
[19:27] <CarlFK> should I report it to lp?
[19:29] <cjwatson> yes please
[19:29] <cjwatson> the bug is that /dev fails to mount, not that stuff about base-installer/section/setup_dev
[19:29] <cjwatson> (so that you get a reasonable title)
[19:29] <CarlFK> package?
[19:29] <cjwatson> base-installer
[19:35] <cjwatson> CarlFK: (it's kind of odd since base-installer hasn't changed since intrepid ...)
[19:35] <cjwatson> bet it's mount --bind vs. mount -o bind
[19:35] <cjwatson> I'll sort it out anyway
[19:36] <CarlFK> is it trying to mount /dev ? (/dev/sda1 is already mounted - might be the problem?)
[19:37] <CIA-2> ubiquity: cjwatson * r2950 ubiquity/ (debian/changelog ubiquity/components/console_setup.py):
[19:37] <CIA-2> ubiquity: Remove seen flag from console-setup/model and console-setup/codeset to
[19:37] <CIA-2> ubiquity: make sure that they don't get misinterpreted as preseeding (LP: #94177).
[19:37] <cjwatson> CarlFK: /dev/sda1 has nothing to do with it
[19:38] <CarlFK> https://bugs.edge.launchpad.net/ubuntu/+source/base-installer/+bug/300426
[19:38] <ubottu> Launchpad bug 300426 in base-installer "alt-install /dev fails to mount" [Undecided,New]
[19:41] <cjwatson> ta
[19:42] <CIA-2> installation-guide: cjwatson * r441 ubuntu/debian/changelog: releasing version 20081113ubuntu1
[20:10] <cr3> how can I configure the preseed on jaunty to prevent it from attempting to fetch packages from updates and security, I'm not sure if that's causing my installation to fail: Nov 20 20:05:01 base-installer: warning: apt update failed: 100
[20:11] <cjwatson> I'm not entirely sure I'm worried about preseeding on jaunty just yet!
[20:11] <cjwatson> but I can only look at the full syslog, not a one-line snippet
[20:12] <cr3> cjwatson: that reminds me, is there a pastebin that allows me to update a file? I don't mind sending it to people.u.c, but just wondering for other people
[20:12] <cjwatson> I generally suggest paste.ubuntu.com unless people have their own favourite
[20:13] <cjwatson> though don't think it allows updates
[20:13] <cr3> cjwatson: yeah, doesn't support uploads
[20:13] <cr3> http://people.ubuntu.com/~cr3/syslog
[20:13] <cjwatson> doesn't support uploads?
[20:13] <cjwatson> oh you mean a file browser button thing?
[20:14] <cr3> cjwatson: oops, I think we got our wires crossed. pastebin.u.c doesn't support uploads and preseed doesn't support updates, does that sound right?
[20:14] <cjwatson> just use http://people.ubuntu.com/~cjwatson/ubuntu-paste
[20:14] <cjwatson> thanks for the syslog
[20:14] <cr3> ok, I think we got our wires straight now. sorry for the confusion
[20:15] <cr3> cjwatson: by the way, your ubuntu-paste is either not executable or not allowed to execute
[20:15] <cjwatson> cr3: the apt update failing is not immediately relevant - the real failure you're running into is bug 300426, which CarlFK mentioned above
[20:15] <ubottu> Launchpad bug 300426 in base-installer "alt-install /dev fails to mount" [Undecided,New] https://launchpad.net/bugs/300426
[20:15] <cjwatson> cr3: executability is your own problem. chmod +x
[20:15] <cjwatson> download it, I don't expect people to run it on rookery :)
[20:15] <cr3> cjwatson: understood, thanks!
[20:16] <cjwatson> (I don't run it there and I don't think it's good to encourage people to run executables out of other people's home directories ...)
[20:17] <cjwatson> by the way, the jaunty-security and jaunty-updates Packages files exist on the main archive - this is a problem with your mirror
[20:17] <cjwatson> easiest approach is just to mirror those pockets
[20:18] <cjwatson> (they're empty, shouldn't take much effort ...)
[20:19] <cr3> cjwatson: the problem is that I'm testing the alternate image, so I don't want anything retrieved from -updates and -security. however, since that won't prevent the installation from continuing and the problem is the mount bug you mentionned, I should be alright on that front
[20:20] <cjwatson> -updates and -security will be empty until release; there is no reproducibility concern with just letting it fetch from there
[20:20] <cjwatson> but you could set up empty pockets locally if that is a concern
[20:21] <cr3> cjwatson: the concern is simply that I'm using the iso image as my mirror, so I'd have to do something special to handle -updates and -security differently at the web server level
[20:22] <cjwatson> yes, I suppose you would
[20:22] <cjwatson> as I've said before, the use case of using an iso image as a mirror is not truly supported and only works by accident
[20:22] <cr3> cjwatson: something for my todo list...
[20:23] <cjwatson> and indeed it may have to be broken entirely in jaunty
[20:23] <cjwatson> there's a sizeable chunk of CD space that could be freed up at the cost of breaking that
[20:24] <cjwatson> there are other things I can do first, but ...
[20:24] <cr3> cjwatson: agreed, and I assume responsibility. if it breaks, I'm confident we'll find a way to continue automating testing somehow
[20:25] <cjwatson> I won't break it unless I have to
[20:25] <cr3> cjwatson: if you plan to fee up that space, please let me know ahead of time as much as you possibly can. I'm not worried about finding a workaround, I'm more worried about not providing test results during my turnaround time
[20:25] <cjwatson> yes
[20:26] <cjwatson> basically the problem is that in order to support this case we have to duplicate some udebs between the initrd and /pool/
[20:26] <CarlFK> cr3: what's wrong with letting it hit security.ubuntu.com ?
[20:27] <cr3> cjwatson: by the way, that mount bug doesn't sound hardware specific and should affect all my installation. does the bug also affect installations directly from the media?
[20:27] <cjwatson> the mount bug is not hardware-specific and affects everyone
[20:27] <cjwatson> it's about first or second on my list
[20:28] <cr3> CarlFK: the problem is that I test the iso image using a network install, so my archive is the mounted image which doesn't contain -updates nor -security
[20:28] <cjwatson> I just want to reproduce it first
[20:29] <CarlFK> cr3: but what is the problem? (given I don't think it is affecting anything)
[20:30] <cr3> CarlFK: there is no problem other than my misunderstanding. I thought "warning: apt update failed: 100" was affecting the installation but it turns out it was the mount problem that cjwatson is working on
[20:30] <CarlFK> ah. carry on :)
[20:42] <cjwatson> something distinctly odd in the innards of busybox here. what is going on
[21:52] <superm1> evand, i didn't realize until I read that foundations team summary that you were doing an SRU for usb-creator to intrepid. are the patches I submitted going to be in that SRU, or just a few other more critical things?
=== TheMuso_ is now known as TheMuso