Hello Adam,
I do it differently (with some advantages, I think).
In your case the boot partition is not a raid partition.
The disadvantage is that the two boot partitions will not be in sync.
I.e. they will boot differently in case you have (maybe even
implicitely, not having noticed) changed anything on your current boot
partition. One example would be a kernel update.
You will have to manually ensure that the two boot
partition remain in sync (except grub.conf, of course).
My idea is as follows:
boot records identical (dd if=/dev/bootdisk of=/dev/other_disk bs=512count=1)
/boot also a raid parition,
Always boot off the first disk
When one disk is broken you need physical access to the disk, then
in case of
hotswap:
exchange broken disk with running system, goto rebuild
no hotswap:
second disk is broken:
exchange broken disk,
boot off first and goto rebuild
first disk is broken:
swap first and second disk, goto second disk broken
rebuild
1) rebuild all raid partitions
2) mirror the bootsector: dd if=/dev/healthy_disk of=/dev/new_disk bs=512 count=1
3) done
Hope this helps,
Chris
Saturday, April 15, 2006, 6:04:49 AM, you wrote:
AC> I used the following a while back and it worked for me....
AC> Setting up GRUB: (assuming you've already installed it)
AC> ------------------------------------------------------------------
AC> # grub
grub>> root (hd0,0)
AC> Filesystem type is ext2fs, partition type 0xfd
grub>> setup (hd0)
AC> Checking if "/boot/grub/stage1" exists... yes
AC> Checking if "/boot/grub/stage2" exists... yes
AC> Checking if "/boot/grub/e2fs_stage1_5" exists... yes
AC> Running "embed /boot/grub/e2fs_stage1_5 (hd0)"... 16 sectors are
AC> embedded.
AC> succeeded
AC> Running "install /boot/grub/stage1 (hd0) (hd0)1+16 p
AC> (hd0,0)/boot/grub/stage2 /boot/grub/grub.conf"... succeeded
AC> Done.
grub>> root (hd1,0)
AC> Filesystem type is ext2fs, partition type 0xfd
grub>> setup (hd1)
AC> Checking if "/boot/grub/stage1" exists... yes
AC> Checking if "/boot/grub/stage2" exists... yes
AC> Checking if "/boot/grub/e2fs_stage1_5" exists... yes
AC> Running "embed /boot/grub/e2fs_stage1_5 (hd1)"... 16 sectors are
AC> embedded.
AC> succeeded
AC> Running "install /boot/grub/stage1 (hd1) (hd1)1+16 p
AC> (hd1,0)/boot/grub/stage2 /boot/grub/grub.conf"... succeeded
AC> Done.
grub>> quit
AC> ------------------------------------------------------------------
AC> Here is how /boot/grub/grub.conf is: (/dev/md0 mounted as /boot)
AC> (Assuming kernel is installed as /boot/bzImage, and RAID1 support
AC> compiled into the kernel).
AC> ------------------------------------------------------------------
AC> # Boot automatically after 30 secs.
AC> timeout 30
AC> # By default, boot the first entry.
AC> default 0
AC> # Fallback to the second entry.
AC> fallback 1
AC> # For booting with disc 0 kernel
AC> title GNU/Linux (hd0,0)
AC> kernel (hd0,0)/bzImage root=/dev/md1
AC> # For booting with disc 1 kernel, if (hd0,0)/bzImage is unreadable
AC> title GNU/Linux (hd1,0)
AC> kernel (hd1,0)/bzImage root=/dev/md1
AC> ------------------------------------------------------------------
AC> -Adam
AC> On 4/14/06, Brian N. Smith <brian (at mark) nuonce.net> wrote:
>> > I can't use the LILO boot loader as it always hangs on boot on my server.
>> > Grub has worked fine but it appears when installing with Grud as the boot
>> > loader there is no raid support, is this true?
>>
>> There are major issues with Grub RAID. I have it on my schedule of things
>> to figure out WTF is going on. With LILO, you simply type /sbin/lilo and it
>> rebuilds the MBR of both Primary & Secondary. Well, with GRUB, there
>> doesn't appear to be an easy way to do it.
>>
>> If someone knows a sure fire way to copy the MBR WITHOUT any problems, I
>> will be happy to quickly publish a new ISO Installer.
>>
>> Thanks,
>> Brian N. Smith
>> NuOnce Networks, Inc.
>>
>>
>>
AC> --
AC> -----------------------------------------------------------------
AC> Shroom.net Donation Based Web Hosting
AC> http://www.shroom.net/
AC> -----------------------------------------------------------------