[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

testing RAID-1 on an Ultra 10



Hi,

I'm experimenting with using RAID-1 on an Ultra 10 workstation with two 9GB
IDE disks, existing as hda and hdd (hdc being a CDROM drive).  The machine
is running Woody with a custom-compiled 2.4.19 kernel, all relevant drivers
compiled in rather than using modules.  I've successfully set up all
partitions on RAID save for one small boot partition on each drive and
verified that I can boot the machine off either.

Next I've been testing booting with one drive removed, in order to simulate
total failure of that disk.  If I remove hdd, there are no problems: the
machine boots fine with degraded arrays.  I then rebuild the arrays and try
again with hda removed.  I get the SILO prompt with no problems, and the
machine boots and mounts the root partition.  What it appears unable to do
from there is to mount any other partitions.  Instead I get errors as
follows:

md: could not lock [dev 03:04], zero-size? Marking faulty.
md: could not import [dev 03:04]!
md: autostart [dev 03:04] failed!

Shortly thereafter, fsck.ext2 failed ("Invalid argument while trying to
open /dev/md4") and drops me to manual repair mode.  Marking hda as
faulty in /etc/raidtab and rebooting has no effect.  I've also tried adding
details of additional md devices at the SILO prompt, again with no luck.

I get the same behaviour if I move the drive hdd to be hdc, ie secondary 
master, but if I make it hda (primary master), the machine starts up with
no problems.  I've had no problems in removing and re-adding partitions
from the raid devices through software (rather than physical
disconnection).

Am I missing something and there's an obvious reason why booting from hdd
in this manner is not going to work?  If not, does anyone have any ideas as
to how to get around the problem?  


FYI:

My silo.conf reads as follows:

partition=2
root=/dev/md2
timeout=100
read-only
image=/boot/vmlinux-2.4.19
label=linux
append=" md=2,/dev/hda2,/dev/hdd2"

while /etc/fstab is

/dev/md2	/	ext2	errors=remount-ro	0	1
/dev/md4	/usr	ext2	defaults		0	2
/dev/md5	/home	ext2	defaults		0	2
/dev/md6	/var	ext2	defaults		0	2
/dev/md7	/data	ext2	defaults		0	2
/dev/md8	none	swap	sw			0	0

Boot partition /boot is manually mounted from either hda1 or hdd1 and
contains the actual silo.conf file (to which /etc/silo.conf is a link).

RAID devices are configured as follows in /etc/raidtab:

raiddev /dev/md2
        raid-level      1
        nr-raid-disks   2
        nr-spare-disks  0
        chunk-size      4
        persistent-superblock 1
        device          /dev/hda2
        raid-disk       0
        device          /dev/hdd2
        raid-disk       1

and similarly for md[4-8]

-- 
--------------- Robin Stevens  <robin.stevens@oucs.ox.ac.uk> -----------------
Oxford University Computing Services ----------- Web: http://www.cynic.org.uk/
------- (+44)(0)1865: 273212 (work) 273275 (fax)  Mobile: 07776 235326 -------



Reply to: