Hey everyone, I was having this exact same problem. At some point an older array was defined, and there were superblocks for md left on one (or more) of the partitions. Even if I zeroed them out (using sudo mdadm --zero-superblock /dev/sda1), the superblock came back after a reboot! It was very strange. Regardless if the partitions were of type "linux" or "linux raid autodetect", I was getting the above errors and problems as people have already described. However, if the partitions were CLEARED completely, the above repeating error did not occur.
SOLUTION FOR ME: In the file: /etc/udev/rules.d/85-mdadm.rules Comment out the only line: SUBSYSTEM=="block", ACTION=="add|change", ENV{ID_FS_TYPE}=="linux_raid*", RUN+="watershed -i udev-mdadm /sbin/mdadm -As" It looks like for some weird reason, udev is trying to be helpful and searching for all devices with md superblocks and assemble any array it might find? Clearly this is useless because mdadm itself already does this. HENCE: It makes sense that we're getting the above error: "md: array md0 already has disks!" (although I'm not sure why it happens over and over again). Anyways, this is what worked for me (thanks to RaceTM for help!) Cheers - hopefully my details will help the developers implement this bug fix. -- mdam software raid fails to start up on reboot https://bugs.launchpad.net/bugs/188392 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs