This is the workaround to boot the system. In busybox manually start raid volumes:
mdadm --run /dev/md0 mdadm --run /dev/md1 mdadm --run /dev/md2 mdadm --run /dev/md.... close busybox with exit system will now boot on next reboot everything is fine because the arry was in state: active, degraded and the faulty member is now detached. If you have lvm the system will not boot on first close of busybox. You can try to close the second busybox (with exit) and then reboot the system. Now lvm should start correctly.