Subject: Re: raidframe
To: None <netbsd-help@netbsd.org>
From: Simon Burge <simonb@wasabisystems.com>
List: netbsd-help
Date: 11/10/2000 12:13:49
Manuel Bouyer wrote:

> I worked around this by splitting the drives, and using a small raid 1 for
> / + swap and a larger one for the rest. This way, after a reboot the root raid
> is dirty but rebuilding it doesn't take too much time.

In my case I did something similar to Manuel - but I have swap on it's
own raid set, and / + everything else on a larger set.  This was, even
the rootfs is clean on a reboot.  I also have my 'b' partition overlaid
64 blocks into the raid set that contains the swap partition:

  #        size   offset     fstype   [fsize bsize   cpg]
    a: 70024752        0       RAID                        # (Cyl.    0 - 69468)
    b:  2096576 70024816       swap                        # (Cyl. 69469*- 71548)
    d:  2096640 70024752       RAID                        # (Cyl. 69469 - 71548)

and have "dumps on wd0b" in my kernel config (although I could more
easily have put a "dp" entry in /etc/fstab).

As of NetBSD 1.5 with alpha and pmax you can load your kernel directly
off the mirrored root filesystem as well.  There's support for vax and
partial support for i386 in -current too.

Simon.
--
Simon Burge                            <simonb@wasabisystems.com>
NetBSD Sales, Support and Service:  http://www.wasabisystems.com/