[03:35] <secureboot> cjwatson: did you have any ideas on the partman error that says "can't have a partition outside of disk"
[07:30] <secureboot_> anyone know how to preseed with the first hard disk available?
[07:30] <secureboot_> i mean, make partman use the first hard disk available
[07:30] <secureboot_> the examples that do that are from dapper, and don't work
[07:35] <secureboot_> d-i partman-auto/disk string /dev/sda will work if i have an sda disk, hda works for hda, but i want something that works for both
[07:35] <secureboot_> any ideas?
[07:55] <secureboot_> i've begun to think there's a bug in partman for PATA disks
[07:56] <secureboot_> the same recipes that work for hdas, don't work for sdas when you have PATA disks
[07:56] <secureboot_> I get "no root filesystem" when I try the same recipes
[07:56] <secureboot_> whereas they work fine on hdas (QEMU, for instance)
[08:03] <cjwatson> first hard disk available> not possible in edgy; use dapper or feisty
[08:03] <cjwatson> well, only possible with grungy scripting
[08:03] <cjwatson> or you could use Kickstart, which should be fine in edgy. See the documentation for clearpart
[08:04] <cjwatson> if you're getting "no root file system" then your recipe is buggy in some subtle way; I'm pretty confident that partman couldn't care less whether it's hda or sda (and much less whether it's PATA) at that level
[08:11] <secureboot_> cjwatson: hmm - i use exactly the same recipe
[08:12] <secureboot_> cjwatson: you were mentioning something 2 days ago about this maybe being related to an error someone found
[08:18] <cjwatson> you were talking about "can't have a partition outside of the disk" which is surely different?
[08:18] <secureboot_> cjwatson: yes, that's true
[08:19] <secureboot_> cjwatson: you also said that recipe looked correct, however
[08:19] <secureboot_> cjwatson: i was wondering if there was just a whole class of problems on PATA disks
[08:22] <secureboot_> http://pastebin.ca/378772 is what i'm using now
[08:25] <cjwatson> providing the disk is big enough, of course
[08:25] <secureboot_> cjwatson: it is
[08:26] <cjwatson> did I already ask you for a DEBCONF_DEBUG=5 log? it would probably also help to add 'set -x' to /bin/autopartition and /bin/perform_recipe
[08:26] <cjwatson> takes quite a while to debug this sort of thing unfortunately
[08:27] <secureboot_> cjwatson: huh - i'll try to do that - how do i add to /bin/autopartition?  where is that /bin?
[08:27] <cjwatson> you have to edit it from tty2 with nano
[08:28] <cjwatson> I would suggest leaving the hostname unpreseeded, and then you'll be able to edit it while the installer is waiting for the answer to that question
[08:28] <cjwatson> then extract /var/log/syslog and /var/log/partman
[08:28] <cjwatson> (after the failure, obviously)
[08:29] <secureboot_> cjwatson: oh - neat.  I'll try to do that then
[08:31] <cjwatson> fairly standard approach to this sort of probblem
[08:31] <cjwatson> it's quite data-intensive and you need the code as a reference to work out what it means
[08:31] <cjwatson> it's definitely useful to add DEBCONF_DEBUG=5 to the kernel boot parameters so that you get a debconf protocol trace in there too
[08:51] <secureboot_> i didn't evne know /var/log/partman existed - i'll be happy to look at that
[09:15] <CIA-4> migration-assistant: evand * r38 migration-assistant/ (10 files in 2 dirs): Fixes for non-English systems.