#1 Software RAID1 with failed disk

[Morton Jonuschat]

Situation:

Trying to migrate a root partition to software raid.
Linux installed on /dev/sda1
Additional disk /dev/sdb installed and partitioned
Software raid constructed, dumped and restored
/dev/sda1 to /dev/sdb1 (RAID1 across /dev/sda1 +
/dev/sdb1, /dev/sda1 marked as failed).

Booted from /dev/sda1 with root=/dev/md0, works fine.
Silo 1.2.5 fails with a message that it can't find out
where to load second stage from when I try to install it.

Upgraded to silo 1.2.6 from debian/unstable. Same
message. Recompiled package with a few print commands
in silo/silo.c.

Found out that the disk which is marked as failed
(/dev/sda1) gets reported with minor/major no. 0. This
falls thru to the default case which leads to above
error message.

Change the major/minor == 0 logic in silo/silo.c a bit
to make it install the bootloader without problems.

Patch is attached.

[Morton Jonuschat]

Patch to make silo skip disks marked as failed in software raid