On Wed, Jun 22, 2005 at 11:52:59AM +0100, David Brownlee wrote:
> On Wed, 22 Jun 2005, Geert Hendrickx wrote:
> 
> >On Wed, Jun 22, 2005 at 06:22:39PM +0800, Cameron Patrick wrote:
> >>Malcolm Herbert wrote:
> >>
> >>>I'm putting together a system with two internal disks in a mirror
> >>>configuration and I'd like to do away with the need to have the extra
> >>>small slice to house the kernel outside the RAID device.
> >>
> >>Erm, doesn't this work anyway without jumping through hoops?  I've
> >>never tried it, but the raidctl man page claims that the bootloader can
> >>read a kernel straight from a raid1 set.
> >
> >He wants to read the kernel directly (not via raidctl) from an
> >individual disk which later in the boot sequence becomes part of a raid
> >set.
> 
> 	That has worked since 1.6.x on i386. Just create a raid set and run
> 	installboot on each components 'a' partition.

Yes, that's what I meant.  And now that I'm re-reading the parent post, I
see I mis-interpreted it.  I thought he (Cameron Patrick) meant there would
be hoop-jumping involved.  Ehrr, double negation. :-)  

GH

-- 
:wq