A couple of months ago my box at home died. I came back home for vacation two days ago and started to check why it doesn't start.. One of the disks had been marked as failed (or acctually two, but only one is dead), I was wondering of removing all the data by re-creating the array and scrapping ~3Tb data.. But when I woke up from a nap this evening I thought that maybe the -f flag works with mdadm as well:
# mdadm -A -f /dev/md0 /dev/sd[acdefgh]1
Did the trick, it resetted the failed status on sdh1