mdadm faulty disk

I have a raid 10 array which reported two failed disks (of 5). So I
added two more. However those two are still marked as spare. As well
I can't remove the failed drives.

# cat /proc/mdstat
Personalities : [raid10]
md0 : active raid10 sdg1[4](S) sdh1[5](S) sdf1[6](F) sde1[7](F)
sdd1[2] sdc1[1] sdb1[0]
781417472 blocks 64K chunks 2 near-copies [4/3] [UUU_]

I don't recall ever having to do something to make mdadm start using a spare.

udev has already removed the device nodes...

#mdadm /dev/md0 -r /dev/sdf1 -r /dev/sde1
mdadm: cannot find /dev/sdf1: No such file or directory

...which is why I assume the not found message. Do I really have to
recreate the device nodes to remove the faulty drives? I don't recall
having to do that before either...
Received on Thu Jun 7 14:44:44 2007