Hi, I recently tried to remove a faulty hard drive from a software
RAID array.

(My log trail is here:
http://groups.google.com/groups?hl=en&lr=lang_en&ie=UTF-8&oe=utf-8&safe=off&threadm=fbef97c.0401182336.14f167f3%40posting.google.com&rnum=2&prev=/groups%3Fq%3Didiotprogrammer%2540yahoo.com%26num%3D100%26hl%3Den%26lr%3Dlang_en%26ie%3DUTF-8%26oe%3Dutf-8%26safe%3Doff%26sa%3DN%26scoring%3Dd
)

I had three drives, hda, hde, and hdg (hde and hdg were on the
controller card; hda was on the primary controller). hda was the
defective drive.

When I replaced hde with a new hd, I discovered voila--not even a boot
prompt!

/md1 boot ext3
swap on /dev/hda2,/dev/he2 and/dev/hdg2 (not RAIDed)
/md3 /  (raid 5) reiser
/md4 /home (raid 5) reiser

Two possible causes: 1)bios allowed boot only from motherboard's
controller's
2)grub was configured to look for /dev/hde1 (on the controller card). 

Here's what I tried:
Possibility 1)I connected the hdg drive as the slave and hda as the
master on the primary mobo controller (that made hdg into hdb). When I
did this, it did not boot at all.

Possibility 2) I booted from the gentoo CD, changed grub so that it
looked for /dev/hdg (and later /dev/hdb). It booted md3 into read only
maintenance mode. I could not perform any permanent write actions and
I couldn't unmount anything, so I pretty much didn't know what to do.

Here are my questions: 
1)Is there a way to have bios boot from a hard drive on a controller
card? Let's assume for a moment that the mbr was located either on the
defective drive (hda) or on hdb.
2)If maintenance mode doesn't allow write-operations or any RAID type
commands, what's the point of it?
3)When I changed the drive that used to be hdg into hdb (slave of
primary controller), why should linux care that the letter name for
the drive has changed?
4)Is it normally necessary to do RAID recovery manually, or this just
something that ought to happen automatically after you boot up for the
first time with the replacement drive?

rj 
Robert nagle www.idiotprogrammer.com