Bug#639830: mdadm: alternative md-device names

[Cristian Ionescu-Idbohrn]

On Wed, 31 Aug 2011, martin f krafft wrote:
>
> Can you add 'set -x' to the top of the hook, remake the initramfs,
> and put it somewhere for me to download (or
> ftp://ftp.madduck.net/incoming)?

Alright.  I see I've drawn too early conclusions :(
Luckly, I've been able to reproduce.  After doing the above, and after
reboot I'm dropped to the busybox-ash shell:

...
Begin: Loading essential drivers ... done.
Begin: Running /scripts/init-premount ... done.
Begin: Mounting root file system ... Begin: Running /scripts/local-top ... Begin: Loadingdone.
cannot initialise MD subsystem (proc/mdstat missing)
[    4.261530] device-mapper: uevent: version 1.0.3
[    4.285475] device-mapper: ioctl: 4.20.0-ioctl (2011-02-02) initialised dm-devel at redhat.com
done.
Begin: Waiting for root file system ... done.
Gave up waiting for root device.  Common problems:
 - Boot args (cat /proc/cmdline)
   - Check rootdelay= (did the system wait long enough?)
   - Check root= (did the system wait for the right device?)
 - Missing modules (cat /proc/modules; ls /dev)
ALERT!  /dev/md0 does not exist.  Dropping to a shell!


BusyBox v1.18.5 (Debian 1:1.18.5-1) built-in shell (ash)
Enter 'help' for a list of built-in commands.

/bin/sh: can't access tty: job control turned off
(initramfs)

Now:

(initramfs) mdadm --assemble --scan
...
mdadm: /dev/md/0 has been started with 2 drives.
...
^D

The boot process continues as expected an I get a login prompt.

Uncommented the '# raid1' line in /etc/initramfs-tools/modules.
Regenerated the /boot/initrd.img-$(uname -r).
After reboot, I get the expected login prompt :)

What the whole thing seems to boil down to (unbootable vs bootable) is
something in 'copy_exec' leading to this:

73,78c73
< ++ dirname /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/md-mod.ko
< + mkdir -p /tmp/mkinitramfs_PyD9Ky//lib/modules/3.0.0-1-686-pae/kernel/drivers/md
< ++ dirname /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/md-mod.ko
< + cp -pL /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/md-mod.ko /tmp/mkinitramfs_PyD9Ky//lib/modules/3.0.0-1-686-pae/kernel/drivers/md
< + '[' n = y ']'
< ++ modinfo -F firmware /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/md-mod.ko
---
> + continue
129,134c124
< ++ dirname /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/raid1.ko
< + mkdir -p /tmp/mkinitramfs_PyD9Ky//lib/modules/3.0.0-1-686-pae/kernel/drivers/md
< ++ dirname /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/raid1.ko
< + cp -pL /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/raid1.ko /tmp/mkinitramfs_PyD9Ky//lib/modules/3.0.0-1-686-pae/kernel/drivers/md
< + '[' n = y ']'
< ++ modinfo -F firmware /lib/modules/3.0.0-1-686-pae/kernel/drivers/md/raid1.ko
---
> + continue

(update-initramfs.01.unbootable.log.gz attached; ftp:ing to
ftp://ftp.madduck.net/incoming failed).
Might make sense looking into the generated
/boot/initrd.img-3.0.0-1-686-pae diff too.  I just didn't get a round
twit, yet.  Let me know if I should.


Cheers,

-- 
Cristian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: update-initramfs.01.unbootable.log.gz
Type: application/octet-stream
Size: 2821 bytes
Desc: 
URL: <http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20110831/4eb523e3/attachment.obj>