[wplug] linux software raid

[Jonathan Billings]

Rick Reynolds wrote:
> Will Linux raid tolerate that?  E.g., the spare drive would probably 
> change its name from /dev/hdb to /dev/hdk (or something similar).  Would 
> the array just figure that out and put me back in business?

So, basically, you want it to fail to a spare on another controller, 
then be able to move the disk, which has already been rebuilt and 
integrated into the array.

I believe if your distro uses raid autodetection, it should just work. 
There is metadata stored on each partition in the persistend superblock 
used that identifies it for the RAID utilities.  You should test this 
first, though.

It might be easier to monitor when the array fails, and cold-swap the 
disk into the system, since it's going to require some downtime anyway 
to swap over the disks, and it'll continue to run while in a degraded 
state until it rebuilds anyway.

-- 
Jonathan Billings <billings at negate.org>