Prev: How to turn off all APT security checking?
Next: What does "Stale NFS file handle" mean?
From: Dr. KERÉKGYÁRTÓ István on 6 Feb 2010 14:20
[sorry if I post multiple times, but my message did not seem to get across,
and moderation isn't supposed to last half a day]

Dear All,

I have a system with a raid1 array. I installed it first with a
degraded raid array, one part of the array being /dev/hda2. After
installation I added (mdadm --manage /dev/md2 --add /dev/sda2) the
second drive, which went nicely, the array was assembled. This already
shows that the array is made up of an IDE and a SATA drive (which
might be unwise in general, I don't know, but I have these too drives
at hand currently).

However, when I reboot, only one of the members of the array are
detected. Depending on which drive I boot from (both are bootable with
identical /boot partitions), either /dev/sda2 or /dev/hda2 is attached
to the array (but not the other one). Looking at dmesg suggests that
first the drive from which the system is booted is detected, then the
md array is assembled, and the other drive is only detected
afterwards.

Is there a way to tell the initramfs in which order to detect the
drives it needs (so that both sda and hda would be detected, with the
appropriate modules loaded, and only after that would the raid array
be assembled)?

Thanks,
I.
 | 
Pages: 1
Prev: How to turn off all APT security checking?
Next: What does "Stale NFS file handle" mean?