Package: initscripts
Version: 2.88dsf-13.1
Hi,
I just had a tough upgrade to squeeze. checkfs is failing and bombing
out on boot.
The reason appears to be related to the ordering of the initscripts
related to cryptdisks, lvm2, and mount.
In my configuration, root is an encrypted disk broken up into lvm2
volumes. This raid volume is unlocked by user password entry on boot.
This machine has several other encrypted disks; the keys to each of
these is found on the filesystem of one of the root array's disks.
In short, whereas the current init setup fires:
checkroot > cryptdisks-early > lvm2 > cryptdisks > checkfs > mount
My configuration would require
cryptdisks-early > lvm2 > checkroot > mountroot > cryptdisks > lvm2 >
checkfs > mount
In Etch and Lenny I didn't have any issues which is partly why I'm a
little confused. Is there anything that can be done to make this part of
boot less brittle?
Has a special runlevel for block devices should be created that runs
commands like these (and mdadm) in a loop, collecting the results of
each run, until nothing has changed before continuing and running a
checkfs that fails been discussed?
-BK
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
