luks related change is prevented boots of system with encryption over raid

Tom London selinux at gmail.com
Wed Oct 29 00:40:51 UTC 2008


On Tue, Oct 28, 2008 at 5:37 PM, Bruno Wolff III <bruno at wolff.to> wrote:
> On Tue, Oct 28, 2008 at 15:55:57 -0700,
>  Tom London <selinux at gmail.com> wrote:
>> On Tue, Oct 28, 2008 at 3:35 PM, Bruno Wolff III <bruno at wolff.to> wrote:
>> > I rebooted this morning after pulling in some changes from last night (a mix of rawhide and koji updates) and my system was not booting.
>> > I so far have noticed that the /dev/mapper/luks-8 naming format has changed and I think that is relating to my problems. I haven't gotten
>> > things totally fixed, but i now have my file systems mount under /mnt/sysimage when rebooting. Hopefully now that the /dev directory is
>> > there a new mkinitrd will clean things up.
>> >
>> > I think using the luks ids in the names is a good idea, but it would have been nice to have something automatically handle the transition.
>> > I haven't seen a big heads up about this and there probably should have been one. It also was a bad thing to change right before the freeze.
>> > So far I have changed /boot/grub/grub.conf and /etc/fstab to use the new names. I'll report back again after hopefully finishing getting
>> > things to work again.
>> >
>> This related?  https://bugzilla.redhat.com/show_bug.cgi?id=468856
>
> After my last try I did notice that it was trying to switch to the new root before asking for my passphrase so that it is likely that the
> problem is the same. I'll look at it again and see about using an older version of plymouth. Though I may have to change the device names
> back if that is something that is only different for rescue mode.
>
The problem described in the BZ above is with mkinitrd, not plymouth.

I just reverted mkinitrd (and nash and libbdevid-python), rebuilt the
initrd and all was OK.

tom
-- 
Tom London




More information about the fedora-test-list mailing list