SnipeDOG: Wow, talk about luck. So usually what I do with a raid5 drive fault is the following:
- Remove the faulty drive: `mdadm --manage /dev/mdx -r /dev/sdy`
- Add the new drive
- Copy the partition table using `sfdisk` or similar utility
- Add the corresponding partition back to the array: `mdadm --manage /dev/mdx -a /dev/sdyz
- Now `/proc/mdstat` should show the recovery process
During this time the array is accessible and a power cycle should not destroy any data, the recovery will just continue