Mailinglist Archive: opensuse (3103 mails)

< Previous Next >
Re: [SLE] Strange raid init sequence
  • From: "Carlos E. R." <robin1.listas@xxxxxxxxxx>
  • Date: Sat, 22 Feb 2003 22:46:34 +0100 (CET)
  • Message-id: <Pine.LNX.4.44.0302222237400.19636-100000@xxxxxxxxxxxxxxxx>

The 03.02.22 at 12:48, Paul Uiterlinden wrote:

> > It doesn't work. I think that at that point initrd is not yet
> > loaded, so that raid1 should have to be in the kernel itself, not
> > as a module. Or disable the kernel trying to mount the raid
> > automatically, I think I saw an option. Or just ignore it.
>
> Strange. I have a RAID0 root partion (/dev/md0, ext3) and it works
> fine. In INITRD_MODULES I have: "aic7xxx jbd ext3 raid0 ide-scsi".

No, no, you misunderstood me: I can mount my raid, that is no problem :-)

The thing is that I'm curious why the kernel tries to mount it very early
on the boot process, but it fails. A bit later, initrd is read, the root
partition (reiserfs) is mounted, and then the raid is successfully mounted,
maybe after reading /etc/fstab (this last detail I'm not sure).

But this means the system is trying twice; I was trying to optimize
the process, and understand why it is happening.

My guess is that, for the first try to succeed, the raid code must not be a
module, but linked directly into the kernel. Probably, this would be
needed to boot from a software raid, which is not my case.

In my case, I would like to disable that first try.

--
Cheers,
Carlos Robinson



< Previous Next >
Follow Ups
References