Re: Etch Software RAID Upgrade Trouble & Suggested Installer Improvements
Russel Coker spoke:
: > The rescue mode of the installer needs a step
: > to activate MD devices. Currently, only the plain
: > disk partitions are visible; that's no help.
:
: Sounds like a reasonable request. Filed a bug report?
:
: Note that it has to be an optional measure.
I will, also for the other packages you have indicated.
Martin F. Krafft said:
: > Well, at the least it should scan for partitions marked as a RAID
: > and prompt the user whether or not to start the set.
:
: Other than the partition type, there's no way to do this, and
: setting the partition type is not always desirable.
:
: Anyway, check out /usr/share/mdadm/startall, which I hope to migrate
: to the udeb after etch. Then you can just call the script and have
: all arrays assembled.
I didn't think that in a recovery situation RAID assembly could
be damaging ... that obviously complicates matters.
I was thinking in the line of the rescue disks that were offered
years ago by Linux distributions, that could be configured to
load a kernel and mount the root file system on disk, thereby
sidestepping the Bermuda triangle of boot loader/kernel/initrd.
That's what I would have needed here.
Anyway, I was already hosed when there were no md devices in
/dev (and I didn't know the majors/minors to mknod them).
Kurt Roeckx enlightened
(after Roberto C. Sanchez' minimal art):
: > Well, at the least it should scan for partitions marked as a RAID
: > and
: > prompt the user whether or not to start the set.
:
: You basicly have to start the software raid configuration tool, and
: then
: press finish. After that it will see the drives. I've filed a bug
: about this some time ago:
: http://bugs.debian.org/391474
I think that's exactly what I was looking for, except I wasn't
aware of that.
Thanks everybody, it looks like there is a good chance that the
next installers will handle this situation better.
Claus
--
Claus Fischer <claus.fischer@clausfischer.com>
http://www.clausfischer.com/
Reply to: