opapo | I am working with dfsbuild and am running into some errors that are inexplicable | 00:00 |
---|---|---|
cjwatson | dfsbuild isn't something I've ever used | 00:00 |
cjwatson | so I'm afraid we may have difficulty supporting it here | 00:01 |
cjwatson | however: please quote exact error messages, preferably with context of where they're appearing | 00:02 |
cjwatson | I should know where "Unknown suite Hardy" would be coming from, but I can't find it just now, and context would probably help | 00:03 |
cjwatson | like I said before, "not an Ubuntu CD" simply literally means that .disk/info doesn't exist on the CD; that's the precise condition for that message being displayed | 00:04 |
CIA-1 | user-setup: cjwatson * r103 ubuntu/ (3 files in 2 dirs): | 01:16 |
CIA-1 | user-setup: Create sambashare group and add the initial user to it (thanks, Mantas | 01:16 |
CIA-1 | user-setup: Kriaučiūnas; LP: #238224). | 01:16 |
CIA-1 | user-setup: cjwatson * r104 ubuntu/ (debian/changelog user-setup-apply): | 01:26 |
CIA-1 | user-setup: Stop creating /etc/aliases, since apparently postfix would rather have a | 01:26 |
CIA-1 | user-setup: blank slate (LP: #228391). | 01:26 |
CIA-1 | user-setup: cjwatson * r105 ubuntu/debian/changelog: releasing version 1.20ubuntu2 | 01:29 |
davmor2 | evand ping | 11:22 |
cjwatson | evand_: I'm trying to get something that resembles an intrepid desktop CD to work with before I poke at your localechooser changes, FWIW | 12:10 |
cjwatson | promoted two packages which should sort it out after a couple of publisher runs | 12:11 |
davmor2 | xivulon: good news for the most part wubi seems to be doing it's job hansomly | 12:15 |
xivulon | davmor2: good to hear for a change :) | 12:54 |
xivulon | thanks a lot | 12:54 |
davmor2 | np's | 12:54 |
davmor2 | evand_: ping | 12:55 |
=== davmor2 is now known as davmor2_dinner | ||
evand_ | davmor2_dinner: pong | 13:46 |
=== evand_ is now known as evand | ||
xivulon | TheMuso any progress on 243226? | 14:05 |
davmor2_dinner | evand: amd 64 bit m-a still broken :( | 14:07 |
=== davmor2_dinner is now known as davmor2 | ||
evand | hrm, ok | 14:10 |
davmor2 | evand bug 218673 | 14:10 |
evand | thanks | 14:11 |
davmor2 | np's | 14:11 |
ilowe | hi guys, I'm having trouble with a preseeded install.... the installer hangs just after the "unable to set title for console-setup-udeb" warning message | 16:42 |
ilowe | any clues? | 16:42 |
evand | you'll need to post /var/log/syslog | 16:44 |
ilowe | from the machine I'm installing on? | 16:45 |
evand | yes, there is a save logs option if you back out to the menu | 16:45 |
ilowe | Hmmm... I can't back out anywhere... my install starts up and I get the blue screen of the installer and the line at the bottom just flashes intermittently... can I get the logs through the console there? | 16:46 |
ilowe | nm... I got it | 16:46 |
cjwatson | that "unable to set title" message is purely cosmetic FYI | 16:46 |
cjwatson | so it'll be something else | 16:46 |
ilowe | cjwatson: yeah... but that's the last message I get in the logs | 16:47 |
cjwatson | you can also press alt-f2 and use nc, or alt-f2, 'anna-install openssh-client-udeb', and use scp | 16:47 |
cjwatson | yeah, just saying it's a totally useless message for the purpose of diagnosis and therefore it is necessary to look at something else :) | 16:47 |
ilowe | cjwatson: fair enough... trying the scp now | 16:47 |
cjwatson | it sounds like a menu item is looping | 16:48 |
ilowe | hmmm... that scp doesn't seem to work | 16:48 |
cjwatson | with what error message? | 16:48 |
ilowe | When I install it I get a bunch of blank lines and then scp is not available at the command line | 16:48 |
cjwatson | might depend how far through the installer it got | 16:48 |
cjwatson | try nc instead | 16:49 |
ilowe | cjwatson: murgh... I have no network on the box | 16:49 |
cjwatson | oh | 16:49 |
cjwatson | can I see the preseed file and try zen diagnostics? | 16:49 |
ilowe | cjwatson: let me pastebin the last lines here... hang on | 16:49 |
cjwatson | (blank out any passwords therein) | 16:49 |
ilowe | cjwatson: argh... no way to grab the output (I'm in a qemu window) | 16:50 |
ilowe | cjwatson: I'll post the preseed file | 16:50 |
ilowe | cjwatson: here it is: http://pastebin.com/d7112bb14 | 16:51 |
cjwatson | what's your kernel command line (via pxe or whatever)? | 16:52 |
cjwatson | is this a netboot install? | 16:52 |
cjwatson | oh, and from which URL did you download the netboot image, if so? | 16:53 |
ilowe | cjwatson: here's my kernel "append" line: file=/cdrom/preseed/asterisk.seed debian-installer/locale=en_CA console-setup/ask_detect=false console-setup/layout=us initrd=/install/initrd.gz ramdisk_size=16384 root=/dev/ram -- | 16:53 |
ilowe | cjwatson: not netboot, I modified an installer CD, edited the isolinux.cfg file and added my own preseed file | 16:53 |
ilowe | cjwatson: The thought occurs: do I need to generate/re-generate md5 sums for those files? | 16:54 |
cjwatson | md5sums aren't yet checked unless you ask for it | 16:54 |
ilowe | cjwatson: ok, cool | 16:54 |
cjwatson | that should be console-setup/layoutcode=us | 16:54 |
cjwatson | (and you can leave out console-setup/ask_detect=false, console-setup/layoutcode=anything implies it | 16:55 |
cjwatson | ) | 16:55 |
ilowe | OK, I'll switch it up | 16:55 |
cjwatson | that might be it - it's hard to say for sure | 16:55 |
ilowe | cjwatson: let me give it a shot... hang on | 16:55 |
* cjwatson tries with intrepid | 16:55 | |
cjwatson | yeah, I think that's it, hangs here too | 16:56 |
ilowe | OK... just re-packaging the ISO now | 16:57 |
cjwatson | might be able to catch that, though it's something of a "don't do that then" :-) | 16:57 |
ilowe | cjwatson: nice! it seems to be working; thanks so much | 16:58 |
ilowe | cjwatson: follow-up question: do you know how I can skip the intro bit where I need to select English and then select my install type from the menu? | 16:59 |
ilowe | cjwatson: hmmm... I spoke too soon | 16:59 |
ilowe | cjwatson: it says it failed to process the file... | 17:00 |
cjwatson | yes, that may just have got you further along | 17:00 |
cjwatson | exact error message please? it's hard to grep for paraphrases | 17:00 |
ilowe | cjwatson: something about /etc/default-release doesn't exist | 17:00 |
ilowe | sorry | 17:00 |
ilowe | It says "Failed to process the preconfiguration file from file:///cdrom/preseed/asterisk.seed. The file may be corrupt" | 17:01 |
ilowe | In the red error screen | 17:01 |
cjwatson | fixed the "unable to set title" message upstream | 17:01 |
ilowe | In the console it printed out the following (bits stripped)... | 17:02 |
ilowe | cat: | 17:02 |
ilowe | /etc/default-release | 17:02 |
ilowe | No such file or directory | 17:02 |
ilowe | Then two DEBUG messages about libc6 and libnewt0.52 not existing | 17:02 |
ilowe | the final message is: | 17:03 |
ilowe | INFO: Menu-item 'file-preseed' selected | 17:03 |
cjwatson | could you acknowledge the error message (press enter) and see if anything further appears on alt-f4? | 17:05 |
cjwatson | it's odd that debconf-set-selections would have exited non-zero without something on stderr | 17:06 |
ilowe | OK, now it says (in alt-f4): | 17:06 |
ilowe | /bin/debconf-set-selections: | 17:06 |
ilowe | shift: 56: | 17:06 |
ilowe | can't shift that many | 17:06 |
cjwatson | blink | 17:07 |
cjwatson | give me a minute :) | 17:07 |
ilowe | I may have found something... I had a trailing comma in my tasksel line | 17:07 |
cjwatson | you did, but that shouldn't cause a parse error | 17:07 |
cjwatson | (feel free to try it, I just think it's unlikely) | 17:08 |
ilowe | Do you know what the "shift" thing is? | 17:08 |
cjwatson | I'm investigating | 17:09 |
cjwatson | I mean, superficially yes, but not why | 17:09 |
ilowe | OK | 17:09 |
ilowe | Yeah, well... that didn't work (as you predicted) | 17:10 |
ilowe | Same shift error message | 17:10 |
cjwatson | ah ok, 56 is just a line number, that makes more sense | 17:11 |
ilowe | Line number in what? In one of the installer scripts? | 17:12 |
cjwatson | now, let's try looking at the version of debconf-set-selections you're actually running ... | 17:12 |
cjwatson | I'm just thinking out loud | 17:12 |
ilowe | how do I get the version of that? | 17:13 |
ilowe | It's whatever is packaged with the 8.04 server release CD | 17:13 |
cjwatson | it's ok, I know what you have | 17:14 |
cjwatson | are you able to put the exact file up on a website somewhere, rather than on a pastebin? | 17:15 |
cjwatson | (again with passwords obscured if you like, but with no other changes to the file if at all possible) | 17:15 |
cjwatson | I just want to check that it's not something odd with things that look like spaces but aren't | 17:15 |
cjwatson | since the file looks ok to me | 17:15 |
ilowe | sure... hang on | 17:15 |
cjwatson | basically what it means is that there is a non-comment line with fewer than three words on it | 17:16 |
ilowe | Check here: http://www.ilowe.net/asterisk.seed | 17:17 |
ilowe | The password is in there but it's bogus anyway; don't worry about it | 17:18 |
cjwatson | got it | 17:18 |
cjwatson | (the file I mean) | 17:18 |
ilowe | :) | 17:18 |
cjwatson | oh wow, that's a fun parser bug | 17:20 |
ilowe | :) | 17:20 |
cjwatson | on the line after "d-i partman/confirm boolean true", you have a line containing only a single space | 17:20 |
cjwatson | remove that space | 17:20 |
ilowe | nice | 17:21 |
ilowe | How come it's so picky about spaces? Reminds me of old school make files | 17:21 |
cjwatson | this is fixed in intrepid | 17:21 |
ilowe | Ah.... is there any value to my running intrepid? How solid is it? | 17:21 |
cjwatson | at least I think it's fixed, just a moment | 17:22 |
cjwatson | (phone) | 17:23 |
ilowe | np | 17:23 |
ilowe | looks like it's working | 17:23 |
cjwatson | mm, yeah, there is still a bug in intrepid; I'll get that fixed | 17:39 |
cjwatson | -for line in $(grep -v '^#' "$file" | sed "s/$CR//g"); do | 17:41 |
cjwatson | +for line in $(grep -v '^#\|^[[:space:]]$' "$file" | sed "s/$CR//g"); do | 17:41 |
cjwatson | in case you care | 17:41 |
cjwatson | actually, there's probably a better way, but anyway | 17:41 |
ilowe | cjwatson: what's the script that's in? | 17:42 |
cjwatson | debconf-set-selections | 17:42 |
cjwatson | it's in the preseed source package | 17:43 |
cjwatson | ok, fix in my source tree, will tidy up and commit later | 17:43 |
cjwatson | got to go, hope that sorts out the rest | 17:43 |
ilowe | cjwatson: will [[:space:]] match multiple spaces? | 17:43 |
cjwatson | good point :-) | 17:43 |
ilowe | :) | 17:43 |
cjwatson | the perils of being in a rush | 17:43 |
ilowe | Also, how can I skip the initial menu on the install? | 17:43 |
cjwatson | re intrepid, it's a little less ropey than it was but I certainly wouldn't advise using it for non-testing/development purposes yet | 17:44 |
ilowe | Thanks a lot for your time... I'll be hanging around while I work on this, hopefully I'll find more bugs :) | 17:44 |
cjwatson | sorry, can you wait for a couple of hours until I get back? my stepson's concert is in 16 minutes at school ... | 17:44 |
cjwatson | if you hang around here I'll answer later | 17:44 |
ilowe | cjwatson: np... catch you later | 17:44 |
ilowe | cjwatson: I filed a bug for Hardy and included a patch: https://bugs.launchpad.net/ubuntu/+source/preseed/+bug/245256 | 18:03 |
=== highvolt1ge is now known as highvoltage | ||
evand | cjwatson: am I correct in believing that we could do away with the initramfs diversion in scripts/install.py (r1549), given triggers? | 20:09 |
cjwatson | evand: I'm not sure that linux-image itself is triggerised | 22:52 |
cjwatson | ilowe: FWIW it's usually considered slightly bad practice to attach a patch that came from somebody else without mentioning that ... | 22:52 |
cjwatson | (I've clarified on the bug) | 22:53 |
cs278 | Would anybody happen to know why I am having to set mirror/country as an argument to d-i? as it will not work inside a preseed file | 23:02 |
cjwatson | cs278: hmm, shouldn't be necessary. Can I see the resulting installer syslog? | 23:09 |
cs278 | this is under gutsy | 23:09 |
cs278 | certain partman arguments seem to require the same treatment too | 23:10 |
cjwatson | that *definitely points to some other problem* | 23:10 |
cjwatson | err | 23:10 |
cjwatson | that *definitely* points to some other problem | 23:10 |
cjwatson | there's absolutely no reason partman should normally need to be preseeded on the kernel command line | 23:11 |
cs278 | I thought not | 23:11 |
ilowe | cjwatson: sorry about that; I meant to mention it but then wasn't sure exactly who you were. My bad. | 23:11 |
cjwatson | can I see the resulting installer syslog demonstrating things going wrong, and the preseed file (with passwords obscured if necessary)? | 23:11 |
cjwatson | ilowe: ah well, never min | 23:11 |
cjwatson | d | 23:11 |
cs278 | cjwatson: yeah, just waiting for it to boot up again | 23:11 |
cjwatson | ilowe: if I'd been in I would have said a bug report wasn't really necessary since it's already fixed in my working tree | 23:11 |
cjwatson | but no worries | 23:11 |
ilowe | cjwatson: fair enough; I saw that you weren't the maintainer of the package so I took a chance that they would want to include it in Hardy. I'll get the hang of it given enough time and leeway :) | 23:12 |
cjwatson | I'm part of the upstream maintainer team | 23:13 |
ilowe | BTW, my install ran all the way through no problems | 23:13 |
cjwatson | oh good | 23:13 |
cjwatson | my inclination is probably not for hardy, since now it would take six months for that to roll out into a point release, and there's a simple workaround | 23:13 |
ilowe | just trying to figure out how to skip the initial language and menu selection now | 23:13 |
ilowe | cjwatson: fair enough... what's the schedule for intrepid like? | 23:14 |
cjwatson | but, once 8.04.1 is done (real soon now!) you're welcome to ask the RM for a second opinion | 23:14 |
cjwatson | intrepid will be October this year | 23:14 |
cjwatson | there should be a timeout setting for the initial menu | 23:14 |
ilowe | cjwatson: I'm not picky about it: I know about the issue now so I'll work around it manually for now; it was a n00b mistake anyway | 23:14 |
cjwatson | 'timeout' in isolinux.cfg I think | 23:15 |
ilowe | cjwatson: funny how we developers always build slightly brittle tools for ourselves :) | 23:15 |
cjwatson | it's 'timeout 0' by default - make it 'timeout 1' and that'll be a 0.1 second delay, which should be close enough | 23:16 |
cjwatson | (0 == infinite) | 23:16 |
cjwatson | brittle> good enough, eh? :-) | 23:16 |
cjwatson | I do think it's a bug not something the user can be "blamed" for | 23:17 |
ilowe | cjwatson: worse is better (if it takes me less time to hack it together that way) | 23:17 |
ilowe | cjwatson: beautiful... the skipping works perfectly: now I have a fully automated install | 23:18 |
* ilowe dances a bit | 23:18 | |
cjwatson | fantastic | 23:19 |
cjwatson | (there are other approaches, you could disable gfxboot and set 'prompt 0' or suchlike - the above is just minimal-modification from what you have) | 23:19 |
ilowe | cjwatson: it's perfect; I don't really care if the screen flashes a bit, I just wanted something that wouldn't require any user interaction | 23:20 |
cs278 | cjwatson: http://tmp.cs278.org/ubuntu/syslog & http://tmp.cs278.org/ubuntu/preseed | 23:23 |
cjwatson | cs278: I'd rather have an example of how it goes wrong *without* those things on the command line | 23:29 |
cs278 | oh sorr | 23:29 |
cs278 | *sorry | 23:29 |
cjwatson | then I can, well, debug it going wrong :) | 23:29 |
cs278 | cjwatson: makes sense, i'm a little fed up with it | 23:31 |
cjwatson | I should be able to figure out what's up | 23:31 |
cjwatson | especially with DEBCONF_DEBUG=5 output | 23:31 |
cs278 | detecting network... | 23:32 |
cjwatson | cs278: I'm going to need to go to bed soon ... any progress? | 23:43 |
cs278 | same urls | 23:43 |
cs278 | just done it | 23:44 |
cs278 | I manually pressed the "enter information manually" forgot I needed that before I could get ssh | 23:44 |
cs278 | but thats where it is stalling | 23:44 |
cjwatson | cs278: the log for the actions performed after reading http://preseed.cs278.org/ubuntu.cfg | 23:47 |
cjwatson | cs278: ... doesn't remotely correspond to what you say the contents of the preseed file are | 23:47 |
cjwatson | cs278: so I'd suggest starting by checking that you really have the right preseed file in that location | 23:48 |
cjwatson | for instance, while reading that preseed file, it's setting a bunch of stuff that looks like it must date from a feisty deployment | 23:48 |
cjwatson | Jul 3 22:32:49 debconf: --> SET apt-setup/local0/repository http://repos.cs278.org feisty main | 23:48 |
cjwatson | for example | 23:48 |
cs278 | :-O | 23:49 |
cs278 | what on earth | 23:49 |
cs278 | I just wgetted it from the box and you are right | 23:49 |
cs278 | its got my old fiesty preseed there | 23:49 |
cs278 | okay somethings gone awry with my apache setup just did a force refresh from firefox and the new script appeared | 23:51 |
cjwatson | caching fun? | 23:51 |
cs278 | seems that way | 23:51 |
cs278 | cheers for your help | 23:51 |
cjwatson | no problem | 23:51 |
cs278 | I'm going to find a dunce hat now | 23:52 |
cjwatson | :-) | 23:52 |
cjwatson | bug 32473 is relevant, and I've just noted that it affects preseed too | 23:53 |
cs278 | one more thing | 23:56 |
cs278 | if I hit escape and go to the menu, and choose download preseed file option | 23:56 |
cs278 | will the changes apply? | 23:56 |
cjwatson | should do though the installer might be in a slightly odd state; I'd probably just start again and have a cup of coffee in the meantime or something | 23:58 |
cs278 | thats a lot coffee ;) | 23:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!