Comment 21 for bug 310928

John Ryan ha scritto:
> The problem all seemed to stem from the fact that when I first re-added
> a disk to the array last night (prior to the successful rebuild), the
> name of the array was changed by the BIOS. After a reboot (or in my
> case, powering off for a day), I then had two different array names and
> neither could be started. I had to use "dmraid -x isw_daifjieafd_Array"
> (from either a live CD or the busybox prompt) to delete the new array,

whit dmraid rc15 you can try dmraid -R to add a new drive into the array, from
man page:

{-R| --rebuild} RAID-set [device-path]
              Rebuild raid array after a drive has failed and a new drive is
added. For Intel chipset based systems, there are two methods in which a
              new drive is added to the system.

              1. Using OROM to identify a new drive
                  During system reboot, enter OROM and mark the new drive as the
rebuild drive.
                  After booting to the OS, use the dmraid command to rebuild.

                  Example: dmraid -R raid_set

              2. Using dmraid to identify a new drive
                  Boot to the OS and use the dmraid command with the new drive
as the second parameter.

                  Example: dmraid -R raid_set /dev/sdc

              3. Using hot spare drive
                  Mark a drive as hot spare using the "dmraid -f isw -S"
command. Then use the dmraid command to start the rebuild.

                  Example: dmraid -R raid_set

> then go into the BIOS and re-add the disk again. Unfortunately, this
> resulted in a new name for the array being generated and so on and so
> forth. This went on until I deleted the array on /dev/sda instead of
> the one on /dev/sdb and as a result ended up stuck with the "Failed"
> array.
>

Anyway I think this is a different issue, intel rebuild is a new features
introduced in rc15 and it is quite experimental.

Cheers,
Giuseppe.