I've got RAIDframe working nicely now, and must say that I'm very
happy with the performance.  Pulling out hot-swap disks and rebuilding
onto spares works great, too.  Another good selling point is the ease
of administration: it doesn't take long to get on top of RAIDframe,
and then it's very comfortable.

However: I was having trouble with mirror sets coming up with "Parity
status: DIRTY" after boot.  After I split things up (following a hint
from Greg Oster, that he posted back in August), and put root alone on
raid0, swap alone on raid1, and the rest of my file systems on raid2,
it turned out that it was only the swap mirror set (raid1) that came
up this way.

Looking at /etc/rc.d/raidframe, I discovered that it will only run
'raidctl -P' on RAID sets that have explicit configuration files in
/etc -- which I don't want to use, because I want the autoconfig to
handle it.  I added code there to run that command on all the sets
that were logged as configured in 'dmesg' output, but that didn't
help: my swap set still came up DIRTY.

Finally, I just put 'raidctl -v -P' statements in /etc/rc.local for
all my RAID sets, which now pronounces raid0 and raid2 clean, but
rebuilds parity on raid1 (where swap is), on every boot.  This does
the trick, so I feel safe now, but I tend to think that it shouldn't
be needed.

Something to do with the way 'swapctl -A' is handled by the kernel?

-tih
-- 
Puritanism -- the haunting fear that someone, somewhere, may be happy.