A previous fix that caused cluster mirrors to reattempt recovery to
    resolve intermittent drive failure, now causes indefinite retries
    during permenent drive failure.
    
    This can slow down the machine dramatically, especially if there are
    several mirrors active on the system that utilize the same failed
    drive.
    
    We now attempt one extra pass.  Further extra passes will require
    a reload of the mirror (as before).