[Wolves] RAID on 750mhz processor?

Andy Smith andy at lug.org.uk
Tue Oct 10 12:40:50 BST 2006


On Tue, Oct 10, 2006 at 09:17:51AM +0100, Mark Ellse wrote:
> Andy Smith wrote:
> >
> >An operating system running on commodity PC hardware often doesn't
> >survive this though, not even SATA or SCSI sometimes, so in these
> >situations all you can do is reboot.  At least it should all boot up
> >again even if the drive is failed.
> OK, I agree that this works. But then how do you put another had disk in 
> and get it to rebuild on the new hard disk. That was the part that I 
> found difficult.

Assume you have a /dev/md0 composed of /dev/sda1 and /dev/sdb1.
/dev/sdb dies.

Fail out /dev/sdb1:

# mdadm /dev/md0 -f /dev/sdb1

/dev/md0 is now running degraded.  Assume you shut down and replace
/dev/sdb, you will have a new /dev/sdb.  Partition it again as
normal, then re-add /dev/sdb1 to the array:

# mdadm /dev/md0 -a /dev/sdb1

Array will now begin resyncing in the background.

$ cat /proc/mdstat

to watch the progress.

I find "sfdisk" useful for copying partition files from other
devices to make them all identical.

Don't forget you need grub installed on every device you will ever
need to boot from, which means both disks in a 2-disk RAID-1.

Cheers,
Andy

-- 
http://strugglers.net/wiki/Xen_hosting -- A Xen VPS hosting hobby
Encrypted mail welcome - keyid 0x604DE5DB
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://mailman.lug.org.uk/pipermail/wolves/attachments/20061010/a7abff26/attachment.bin


More information about the Wolves mailing list