If there are any linux raid partitions (type fd) for which mkraid has not
been performed yet, the scripts invoked by the 'raidstart /dev/md0' command
in the rescue environment crash in isys.py line 37 with a complaint:
ValueError: bad md magic in device
This in spite of the fact that the uninitialized type fd partitions are not
associated with md0.
This prevents access to md0.
Instead the scripts should perform initialization for partitions that do
have valid raid superblocks, and ignore partitions that do not, but still
invoke the raidstart program with provided arguments.
This bug exists in RH 6.x rescue environment too.
Should be fixed.