=== johanbr_ is now known as johanbr [09:58] Hi, got a question about luks-support in 9.10 [09:58] seems to have been altered [09:58] I got a laptop with luks-encrypted partitions. While booting with 9.04's kernel, decrypting works [09:58] doesn't work when booting with the new kernel [09:58] any ideas? [10:08] strange, it works for mine nicely [10:09] what happens when booting 9.10's kernel? [10:09] could it be that cryptsetup somehow disappeared and your initrd doesn't have it? [10:11] hmm [10:11] probably my initramfs is faulty [10:12] but how did this happen? [10:12] and how do I restore it? [10:13] wait a sec. how is your setup like? [10:13] is root on a luks partition? [10:13] yes it is [10:13] ./boot is not [10:13] ah [10:13] similar to mine thne [10:13] then* [10:13] ./ and ./home are [10:13] lvm? [10:13] no lvm [10:13] ah [10:13] so they're separate [10:14] okay, do you have your jaunty kernels around with the same installation, or is your karmic and jaunty installation separate? [10:14] (and you should leave the leading ".". "." stands for "current directory") [10:16] I cant leave . [10:16] at least in IRC [10:16] try starting a chat message with a slash [10:16] :-) [10:16] yes, it works with my jaunty kernel [10:17] I just saw some issues with ldconfi [10:17] g [10:17] /blah [10:17] ;-) [10:17] try either // or / / [10:18] well boot using your jaunty kernel and run sudo update-initramfs -u -k all [10:18] see if that works [10:18] I fixed some library-problems, maybe this is already it. Afterwards, I'll try your suggestion (currently rebooting) [10:19] good luck [10:19] at least I can now use vim again... [10:20] while updating, ldconfig said, my libavahi* and libgpg*-files were empty [10:20] heh [10:20] soudns like your isntallation's kinda knocked up [10:20] yeah [10:20] i'd install debsums and run a check [10:20] debsums? never heard of it, i'll try [10:20] it's good for detecting files which differ from what the's installed [10:20] you wouldn't guess what happened to my fathers laptop [10:21] what? [10:21] tossed it out the window and it survived? =p [10:21] /boot was full. But I couldn't remove old kernels, because dpkg --configure -a had no disk space on /boot [10:21] ... [10:21] and there was no "resume"-optin [10:21] the upgrade process just broke [10:23] how annoying [10:24] ah [10:24] you could dpkg --configure -a again [10:24] do you have a cipher= option specified in /etc/crypttab? [10:24] it'll continue from where it left off [10:24] no i don't [10:24] I see, maybe this might be an issue [10:24] cryptostuff /dev/sda2 none luks [10:24] that's the only line there [10:24] aside from the comment [10:25] I see. mayby I'll try to remove that option [10:26] it shouldn't cause issues imo [10:26] the problem is most probably because dpkg failed [10:26] yeah [10:26] make sure dpkg completes whatever it's doing [10:26] update-initramfs found the crypto devices [10:26] give /boot more space or whatever [10:26] let's reboot [10:26] I thought 200MiB are enough for /boot [10:27] but jaunty had a lot of kernel updates [10:27] /dev/sda1 510M 126M 359M 26% /boot [10:27] I see [10:27] 200MiB is kinda cutting it close [10:27] you can purge some of your kernels [10:27] I did [10:27] the older ones [10:27] ah [10:27] but... [10:27] I couldn't do via dpkg or aptitude [10:28] since i had to do dpkg --configure -a first [10:28] and I couldn't execute that because of low disk space [10:28] so I had to remove them via 'rm' [10:28] that's what bugging me [10:28] lol [10:28] i see [10:28] rm, dpkg --configure -a, then purge [10:29] yeah [10:29] exactly [10:32] from SLES, I know `rpm -qf ` and rpm tells, to which pakage this file belongs to [10:32] very handy [10:33] <_maks> dpkg -S file [10:35] ah [10:35] gotto remember that [10:35] or dlocate [10:35] dlocate is a very much faster dpkg -S [10:35] rpm is extremely slow [10:36] both dlocate and dpkg are probably faster [10:36] heh [10:36] well dpkg -S takes quite some time [10:36] but dlocate is almost instantaneous [10:36] it seems dlocate can also do md5sums check [10:36] I guess dlocate builds some sorta hash map [10:36] s [10:36] it does [10:36] no actually it says it keeps a text dump [10:36] o.O [10:37] it says it keeps a text dump from dpkg [10:37] dlocate -l understands regex, hooray! [10:37] and just uses grep [10:37] lol [10:37] $ wc -l dlocatedb [10:37] 408412 dlocatedb [10:37] my my, that's a lot of lines. xD [10:38] grep is fast then :-P [10:39] I had some issues using grep on AIX, but that's another story [10:39] hyperair: cryptsetup works, thanks!! [10:39] heh [10:39] np [10:40] in our company we use SLES, AIX, Solaris on servers. on AIX and Solaris there is only ksh88 available. bah! [10:43] by the way, hyperair, have you seen that recent picture of Linus Torvalds in front of a win7-store? [10:44] http://tr.im/Eqan [10:51] bmhm: i did, yes. [10:51] it was hilarious [10:59] hyperair: can you help me with another LUKS-Issue? [10:59] what's up? [11:00] as I said... no LVM. I am trying to derive keys from root. Works with swap, but won't work with home (both deriving from root) [11:00] what do you mean derive keys from root? [11:01] means using /lib/cryptsetup/scripts/decrypt_derived as option [11:02] hmm [11:02] never done anything of that sort O_o [11:02] what does it make your crypttab look like? [11:03] erhm wait a mo [11:04] hyperair: http://pastebin.ca/1660772 [11:17] bmhm: how do you create your cryptsetup volumes? [11:20] it's been some time since I created them [11:22] heh [11:24] dunno, does it really matter? [11:25] hyperair: it works when I remove the script-part from the /home-line. I just have to type the password twice [11:25] hmm [11:25] is there documentation on this script? [11:30] hyperair: just found the issue [11:30] to use a derrived key, it has to be added to the luks header first [11:30] ? [11:30] luks header? [11:30] what do you mean? [11:30] the partition header created by luks [11:31] it has to be added via cryptsetup luksAddKey [11:32] i don't understand [11:32] dion't you mean luksFormat? [11:33] oh i get it [11:33] i understand now. [11:33] % /lib/cryptsetup/scripts/decrypt_derived crypt-part > /mnt/crypt-part/key-file [11:34] 2.% cryptsetup luksAddKey /dev/sda7 /mnt/crypt-part/key-file [11:34] aha. [11:34] yeah [11:34] i took it from german ubuntuusers-wiki [11:34] ah [11:34] well i faced that problem at first, so i used lvm instead [11:34] i don't regret it, since lvm makes it extremely easy to resize partitions. [11:34] or create new ones [11:35] resizing cryptsetup volumes are kinda.. @_@ [11:35] Yeah, true [11:35] but when I created my luks partitions about 1.5 years ago, there was no appropriate tutorial [11:36] 1.5 years? [11:36] that's about the same time i created mine [11:36] there were cryptsetup/lvm tutorials lying around [11:37] well... [11:37] it's a good thing i used lvm anyway, i learnt loads about lvm (and use it for sg.releases.ubuntu.com now) [11:37] true [11:37] LVM really is a good thing to have [11:37] yep [11:38] btrfs will have some interesting features as well, which will interact with LVM [11:38] hmm [11:38] is it? [11:38] interact with LVM? i thought it was going to provide all kinds of LVM which will make LVM rather useless for it [11:39] ah perhaps that's what I meant [11:40] this is where I get my information from: [11:40] http://www.h-online.com/open/features/The-Btrfs-file-system-746597.html [11:41] it's a translation of the german site "heise.de", which is the publisher of the best-known german computer magazine (called "c't"). [11:41] btrfs has also raid built in. I wonder what happens to md, [11:47] it'll be around for legacy file systems i suppose [11:50] do you know "the h"? great page imho [11:56] nope [11:56] what's that [12:06] the page i just posted an article from [12:09] well anyway. got to go [12:09] thanks for helping === maco_ is now known as maco === maco_ is now known as maco === nightwis- is now known as nightwish === kees__ is now known as kees === BenC2 is now known as BenC === maco_ is now known as maco [20:42] apw: should I modify build-mkschroot to accept lucid as a valid $SUITE, too? [20:42] ugh, this is going to require kludging in a newer debootstrap, too [20:43] ...which of course doesn't exist yet. === SEJeff_ is now known as SEJeff [23:20] lool: have you had any complete kernel panics on your 301 in karmic, particularly recently?