installing new kernel (RAID-related)

[steve%prd.co.uk@localhost (Steve Blinkhorn)]

Having successfully upgraded a local machine to 9.2 from 7.0 (though I
still can't get X working), I've moved to the trickier task of
remotely upgrading a headless server in a data centre from 8.0_RC1.  I
have a remote management console connection on an out-of-band IP
address.

I moved the old kernel to netbsd-8.0_RC1 and the new one to /netbsd.
Its permissions are set to 755 (I tried 744).   But when I reboot it
consistently boots the old kernel.  If I drop down to the boot prompt
and try hd0a:/netbsd the same happens.

Then I reaslised what was happening: I have my disk space, including
the root partition, managed by RAIDframe, but for safety I have a small
complete 8.0 installation on each of my coupled hard disks.   So I
copied the 9.2 kernel onto wd0a, and now it boots into 9.2, and mount
shows the RAID file structure.

So was I just ill-advised to include the root partition in the
RAIDframe setup?  Or is there something special I should do to get the
system to boot from where I expected it to?  

-- 
Steve Blinkhorn <steve%prd.co.uk@localhost>