[09:26] <CIA-4> debian-installer: cjwatson * r997 intrepid-proposed/ (3 files in 2 dirs): Move mainline architectures to 2.6.27-14 kernels.
[09:42] <CIA-4> debian-installer: cjwatson * r998 intrepid-proposed/debian/changelog: releasing version 20080522ubuntu29
[10:13] <CIA-4> localechooser: cjwatson * r147 ubuntu/ (16 files in 3 dirs): merge from Debian 2.11
[10:16] <CIA-4> localechooser: cjwatson * r148 ubuntu/debian/changelog: releasing version 2.11ubuntu1
[10:32] <wolferz> Is there a way to "undo"  something synaptic did?
[10:34] <davmor2> wolferz: #ubuntu is the general channel for help this is a dedicated channel for the installer process
[10:35] <wolferz> ahh, I apologize, I took install to be installations of packages as well as the actual distro
[10:38] <davmor2> wolferz: No sorry.  #ubuntu will help though you may wish to point out what you did too though
[12:09] <corecode> hey
[12:09] <corecode> cjwatson: you around?
[12:10] <cjwatson> yes, though going out for lunch shortly. On IRC it's best to just talk and not mind if you don't get a response until hours later ...
[12:10] <corecode> yah i know
[12:10] <corecode> d-i partman-auto/method string lvm
[12:10] <corecode> d-i partman-auto-lvm/guided_size string 20GB
[12:10] <corecode> d-i partman-auto/choose_recipe select atomic
[12:11] <corecode> that's the relevant portion
[12:11] <corecode> doesn't change the size of / though, / will use up all space on the volume
[12:11] <cjwatson> ok, the preseeding looks correct. Can I see syslog+partman with DEBCONF_DEBUG=developer please?
[12:12] <corecode> i'll try that.  will the log be preserved after installing?
[12:12] <corecode> or do i have to copy it out before rebooting
[12:12] <cjwatson> note that DEBCONF_DEBUG=developer will include *all* debconf activity including passwords, so I recommend modifying your preseed to use innocuous values
[12:12] <cjwatson> the logs will be preserved in the /var/log/installer/ directoyr
[12:12] <cjwatson> directory
[12:12] <corecode> okay
[12:12] <corecode> thanks
[12:12] <corecode> will take a couple of minutes to install though
[12:12] <corecode> also, auto-install/defaultroot doesn't work for me
[12:13] <corecode> seems the installer doesn't run the autoinstall part
[12:13] <corecode> what i mean is, i can't set url=hostname to make it install.
[12:13] <corecode> but that's a different thing
[12:13] <cjwatson> auto-install doesn't work in Ubuntu
[12:13] <corecode> i see
[12:14] <cjwatson> just set the network stuff on the kernel command line
[12:14] <corecode> yah, works sufficiently well
[13:23] <corecode> http://documents.epfl.ch/users/s/sc/schubert/www/partman
[13:23] <corecode> http://documents.epfl.ch/users/s/sc/schubert/www/syslog
[15:23] <CIA-4> usb-creator: evand * r100 future/ (4 files in 3 dirs):
[15:23] <CIA-4> usb-creator: Manage mountpoints. Unmount any devices mounted by usb-creator on failure or exit.
[15:23] <CIA-4> usb-creator: Once again check for sufficient free space on the target device and update the frontend accordingly.
[15:26] <cjwatson> corecode: I need you to run the installer through a little further before extracting the logs; you seem to have given me logs from right after the preseed file is read - I need to see what the partitioner does too
[15:27] <corecode> that's all i have
[15:27] <corecode> the full installation
[15:28] <corecode> syslog covers from boot to reboot
[15:28] <corecode> cjwatson: do you need something else?
[15:34] <cjwatson> oh, sorry, I think the download got truncated
[15:52] <corecode> it is 8 mb large
[15:58] <cjwatson> I did curl blah | less and obviously waited too long before telling less to go to end of file
[16:02] <corecode> nevermind :)
[17:40] <CIA-4> partman-auto-lvm: cjwatson * r220 ubuntu/ (debian/changelog perform_recipe_by_lvm): Make partman-auto-lvm/guided_size preseedable.
[17:41] <cjwatson> corecode: so, err, yeah, this was a bug, sorry.
[17:42] <cjwatson> corecode: with a bit of creativity, you might be able to use preseed/early_command to bodge the fix from http://bazaar.launchpad.net/~ubuntu-core-dev/partman-auto-lvm/ubuntu/revision/220 into the running installer
[17:42] <cjwatson> the path to the relevant file in the installer environment is /bin/perform_recipe_by_lvm
[17:43] <corecode> i see
[17:43] <corecode> will you roll a new installer at some point?
[17:44] <corecode> i mean, one that gets picked up by the netinstaller
[17:44] <cjwatson> not for jaunty
[17:45] <cjwatson> well, probably not; I'm not sure I can justify this as a release-critical thing that justifies a stable update
[17:45] <cjwatson> we have quite a high bar for that kind of change
[17:45] <corecode> oh
[17:47] <cjwatson> it'll be in karmic of course
[17:49] <cjwatson> should be possible to grab a copy of http://bazaar.launchpad.net/%7Eubuntu-core-dev/partman-auto-lvm/ubuntu/download/cjwatson%40canonical.com-20090505164031-xv31hanv1w3fm4qr/perform_recipe_by_lv-20061127181313-x5wza1qumbsi91v1-12/perform_recipe_by_lvm, and get a preseed script to stick that into the installer at run-time
[17:56] <corecode> not such an easy thing i guess
[17:56] <corecode> well maybe
[17:56] <corecode> i could just preseed-fetch the replacement
[17:59] <cjwatson> yeah, should be straightforward enough
[17:59] <cjwatson> or you have wget :)
[17:59] <corecode> so that goes into /bin?
[18:00] <cjwatson> yeah
[18:01] <corecode> do you have a clever way to specify "first hdd" instead of telling the system /dev/sda?
[18:01] <corecode> (unrelated)
[18:03] <cjwatson> define first :-)
[18:04] <cjwatson> in general, disk probe order isn't well-defined
[18:04] <corecode> the one that would be booted from, ideally
[18:04] <cjwatson> mm, it would be awfully nice if Linux had that information
[18:04] <cjwatson> (that's serious, not sarcastic)
[18:04] <corecode> yah
[18:05] <corecode> well, i guess i'm more concerned with /dev/sda vs /dev/hda
[18:05] <cjwatson> if there's only one disk in your system, you can just leave partman-auto/disk unset
[18:06] <cjwatson> if you have more than one, then you can do something like this: d-i partman/early_command string debconf-set partman-auto/disk "$(list-devices disk | head -n1)"
[18:06] <corecode> yah, i think we have two
[18:06] <cjwatson> (that's untested)
[18:06] <corecode> ah, list-devices
[18:07] <corecode> thanks
[18:08] <corecode> now let's see whether that works
[18:11] <corecode> hm.  doesn't seem that that worked
[18:12] <corecode> could it be that preseed/early_command is run too early?
[18:12] <corecode> so that my replacement gets overwritten?
[18:12] <cjwatson> partman/early_command, not preseed/early_command - that wasn't a typo :)
[18:13] <cjwatson> partman/early_command is a special thing that runs after disks are detected but just before the partitioner starts
[18:13] <corecode> yah, i'm talking about the script replacement
[18:13] <cjwatson> oh, right
[18:13] <cjwatson> uh, yeah, it could well be - try doing that in partman/early_command as well
[18:13] <cjwatson> menu-item-numbers.txt says:
[18:13] <cjwatson> 2100 network-preseed
[18:14] <cjwatson> 2300 [...] download-installer
[18:14] <cjwatson> so indeed
[18:21] <corecode> hm.
[18:21] <corecode> doesn't seem to work
[18:22] <corecode> ah my fault
[18:25] <corecode> nice!
[18:25] <corecode> works
[18:27] <cjwatson> oh good
[18:39] <corecode> thanks!
[18:59] <SiDi> Hello folks
[18:59] <SiDi> I'd like to know what happens exactly at the "Analysing the mirror" step of the installer
[19:00] <SiDi> its stuck at 82% with that phrase, 3 times out of 4 so far, installing Xubuntu 9.04 in a vm
[20:16] <cjwatson> SiDi: the easiest way to tell would be to get /var/log/syslog from the installer; you can extract it by switching to tty2, running 'anna-install openssh-client-udeb', and using scp to copy it out
[20:16] <cjwatson> SiDi: the installer doesn't have any messages that read "Analysing the mirror", as far as I can see - are you translating this back to English from another language?
[20:17] <cjwatson> "Scanning the mirror...", perhaps?
[20:18] <cjwatson> it's basically doing an 'apt-get update' there, although it's supposed to tweak the timeout down to be reasonable; it could be that the mirror is responding for you but is just very slow
[20:22] <SiDi> cjwatson: yeh
[20:22] <SiDi> Scanning probably
[20:23] <SiDi> i'm behind a firewall, but i didnt have the problem when i was installing 8.10 vms, behind the same firewall
[20:23] <cjwatson> I wouldn't mind seeing the syslog, if possible
[20:23] <SiDi> cjwatson: the vm died :P i'll make another and send you the syslog, sec
[20:23] <cjwatson> could be a regression of some kind, or it could just be "normal" slowness
[20:23] <SiDi> i got an horrible bandwidth
[20:24] <SiDi> but i dont understand why it doesnt drop on that
[20:24] <SiDi> cause obviously it wouldnt hang for more than 15 mins if i didnt have the internet
[20:24] <cjwatson> if it's getting anything, even at a very low rate, it'll keep going
[20:25] <cjwatson> if you have no (working) internet access at all, it ought to give up quickly
[20:25] <cjwatson> it's a pretty tough judgement call when it is getting something
[20:25] <SiDi> I see
[20:25] <SiDi> i should disable eth0 from the VM before installing then
[20:25] <SiDi> maybe you could add a 'cancel' button for this step too
[20:25] <SiDi> i saw there was one for when it downloads the locales
[20:27] <SiDi> Do you still want a syslog ?
[20:35] <cjwatson> yes please
[20:35] <cjwatson> I thought there was a cancel button already
[20:35] <cjwatson> are you installing from the alternate CD or the desktop CD?
[20:35] <cjwatson> hmm
[20:35] <cjwatson>   * Disable option to cancel progress bar as there are issues with it.
[20:36] <cjwatson> helpful referencing in that changelog entry
[20:36] <cjwatson> # Option to cancel an action does not currently work:
[20:36] <cjwatson> # - application does not seem to always react to a cancel signal
[20:36] <cjwatson> # - debconf-apt-progress does not exit with code 30 when cancelled
[20:36] <cjwatson> # - even if it did, the exit code would be mangled by in-target
[20:36] <cjwatson> # See also thread http://lists.debian.org/debian-boot/2008/01/msg00094.html
[20:36] <SiDi> Desktop, Xubuntu
[20:37] <SiDi> i'll do this asap then
[20:37] <cjwatson> nah, don't bother with the syslog now, if it's just for the absence of the cancel button
[20:38] <SiDi> i got another VM running and vbox doesn't want me to run another VM since i'm using VT-x
[20:38] <SiDi> okay then :p
[20:38] <cjwatson> as that thread notes, there are some underlying problems that mean we can't offer it yet
[20:38] <cjwatson> unfortunately
[20:40] <cjwatson> bug 172879, which I'll target for karmic
[20:43] <SiDi> Great :)
[20:43] <SiDi> wow, its been there for a while
[20:43] <cjwatson> sure, most non-trivial bugs have
[20:44] <CIA-4> console-setup: cjwatson * r104 ubuntu/ (29 files in 6 dirs): merge from Debian 1.33
[20:51] <CIA-4> console-setup: cjwatson * r105 ubuntu/debian/ (changelog config.proto): merge from Debian 1.34
[21:23] <CIA-4> console-setup: cjwatson * r106 ubuntu/debian/ (console-setup.initramfs-hook console-setup.initramfs-top): update for boottime.kmap.gz -> cached.kmap.gz change
[21:26] <CIA-4> console-setup: cjwatson * r107 ubuntu/debian/changelog: releasing version 1.34ubuntu1
[23:27] <SiDi> Good bye