on 5/11/00 1:11 PM, James Manning at [EMAIL PROTECTED] wrote:
> I'd think that, but he's still not put out the /proc/mdstat I asked for
> multiple times
-------
on 5/10/00 2:13 PM in response to Theo Van Dinter at [EMAIL PROTECTED]:
> what does "cat /proc/mdstat" say? do the drives appear in there somewhere?
cat /proc/mdstat
Personalities : [raid0]
read_ahead 1024 sectors
md0 : active raid0 hdh1[1] hdg1[0] 19806976 blocks 16k chunks
unused devices: <none>
-------
> and the dmesg output he showed didn't have hdl involved
> in md0 at all.
Duh! That's exactly my finding as well, and the source of the confusion,
which is why I don't see why it is trying to claim this drive.
> I don't honestly believe hdl, if it even exists, is even
> remotely involved in s/w raid.
Double-Duh! That's my point. First of all, hdl, along with hdj, and hdk,
EXIST. They are attached to the IDE controller (HPT366), and they show up
during the initial kernel identification phase, where they identified
themselves by the descriptors I looked for.
Webmin, an administration tool, properly finds, and displays these drives,
but declares them as being in use by /dev/md0 (which they should not be).
Webmin is NOT the cause of this, as the system behaves the same way with
webmin disabled.
Hence, since this is obviously getting everyone to scratch their heads, I am
reinstalling the system from scratch, recompiling the kernel with IDE and
RAID patches, and trying all of this again.
I will report back when that is completed.
Harry
