Cannot configure bootable RAID on redhat!! PLEASE help!
pub...@rlnt.net
Hi there,
I'm trying to install RedHat 7.0 on a HP NetServer E 50 with dual
4.2Gb SCSI drives, and configure the two drives as a RAID 1 array.
I've done a full install about eight times now, and have come pretty
close. The docs didn't always seem to reflect what I was experiencing
... :(
Anyway, now I'm stuck, so near yet so far (beginning another fresh
install as I write). Please help!
The bottom line is that I can reboot with the RAID mounted but then
the system drops me to a shell and hasn't mounted anything else, not
even /proc ... ? See below for the full story.
I could REALLY use some help! Three days on this now!
Thanks,
Nick
Here it is:
Following the instructions at
http://redhat.com/support/docs/tips/raid/RAID-4.html and at
http://redhat.com/support/docs/tips/raid/RAID-2.35.html, which seem to
fill in each other's blanks mostly, if not completely, I have:
sda1 23Mb /boot Linux native
sda5 203Mb Linux swap
sda6 3835Mb Raid000 Linux RAID
sdb1 203Mb / Linux native
sdb5 23Mb /boot2 Linux native
sdb6 3835Mb Raid001 Linux RAID
when I leave the Disk Druid partition screen.
In LILO Configuration I install the bootloader in the MBR since I plan
to move the contents of /boot to /boot2 later (right?).
In the next screen it says that /dev/sdb1 (only) will be bootable as
'linux' ...
I continue with the installation, and as I understand it, all the
packages and stuff are installed in / on /dev/sdb1 (right?)
At the end, I reboot. All good. System restarts and I log on.
root# df -k
/dev/sdb1 202220 /
/dev/sda1 23302 /boot
/dev/sdb5 23302 /boot2
so I try to follow the instructions:
root# cat /etc/raidtab
raiddev 1
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
chunk-size 4
device /dev/sda6
raid-disk 0
device /dev/sdb6
raid-disk 1
root# mkraid --really-force /dev/md0
[ ... ]
sdb6 [events: 00000001](write) sdb6's sb offset: 3927744
sda6 [events: 00000001](write) sda6's sb offset: 3927744
.
root# mkfs /dev/md0
[ ... ]
writing superblocks and filesystem accounting information : done
root# mkdir /raidtmp
root# mount /dev/md0 /raidtmp
nr_blocks changed to 32 (blocksize 4096, j694496, max_blocks 981936)
^----is that okay?
Anyway, I can see /raidtmp mounted on /dev/md0 in df, and I can chdir
to it ..., so I:
root# tar cf - ./bin ./dev ./etc ./home ./lib ./misc ./mnt ./root
/sbin ./tmp ./usr ./var | (cd /raidtemp; tar xf -)
tar: ./dev/log: socket ignored
root# mkinitrd /boot/initrd-raid1.img --with raid1 2.2.16-22
root# cat /etc/lilo.conf
prompt
timeout=50
message=/boot/message
linear
default=linux
image=/boot/vmlinux-2.2.16-22
label=linux
initrd=/boot/initrd-raid1.img
read-only
root=/dev/md0
root# /sbin/lilo
Added linux *
root# cp /etc/lilo.conf /raidtmp/etc/
cp: overwrite '/raidtmp/etc/lilo.conf'? y
So now the last thing I need to do is edit /raidtmp/etc/fstab but the
instructions say:
"edit the "/(raid_mount_point)/etc/fstab" and change "/" =
(raid_device) "
and it's just not clear to me what line I am supposed to edit ...
I have:
LABEL=/ / ext2 defaults 1 1
is the only line that references / and it does not offer a device ...
???
Should I add a line?
I changed it to
/dev/md0 / ext2 defaults 1 1
I reboot. The shutdown all seems good, I see several messages about
the RAID turning off and shutting down that are all healthy.
then ... FAILURE!!!
"The superblock could not be read or does not describe a correct ext2
file system .... "
And I get "dropped to a shell"
There:
Repair Filesystem 1 # df
Filesystem 1-k blocks Used Available Use%
Mounted on
/dev/md0 3866000 199504 3470112 6% /
df: /proc: No such file or directory
At the shell I can edit /etc/fstab and I set the line back to:
LABEL=/ / ext2 defaults 1 1
I reboot and the df output is the same, but now I can't edit any
files!!! They are all read only and there's now swap space to open
them for writing!
So the entire last three hours' work seems to be down the tubes again!
Anybody, please help!