Linux mdadm does not assemble array, but recreation of array does it

Maybe i'm not very clear in the title. When i'm trying to assemble my raid1 array with mdadm:

sudo mdadm --assemble /dev/md0 /dev/sdc /dev/sdd

It tells me that

mdadm: Cannot assemble mbr metadata on /dev/sdc
mdadm: /dev/sdc has no superblock - assembly aborted

If i reorder devices in command:

sudo mdadm --assemble /dev/md0 /dev/sdd /dev/sdc

It tells the same for sdd:

mdadm: Cannot assemble mbr metadata on /dev/sdd
mdadm: /dev/sdd has no superblock - assembly aborted

Here's some info about drives:

➜ ~ sudo mdadm --misc -E /dev/sdc
/dev/sdc: MBR Magic : aa55
Partition[0] : 3907029167 sectors at 1 (type ee)
➜ ~ sudo mdadm --misc -E /dev/sdd
/dev/sdd: MBR Magic : aa55
Partition[0] : 3907029167 sectors at 1 (type ee)

But! When i'm recreating array with

➜ ~ sudo mdadm --create /dev/md0 -n 2 -l 1 /dev/sdc /dev/sdd
mdadm: /dev/sdc appears to be part of a raid array: level=raid0 devices=0 ctime=Thu Jan 1 07:00:00 1970
mdadm: partition table exists on /dev/sdc but will be lost or meaningless after creating array
mdadm: Note: this array has metadata at the start and may not be suitable as a boot device. If you plan to store '/boot' on this device please ensure that your boot-loader understands md/v1.x metadata, or use --metadata=0.90
mdadm: /dev/sdd appears to be part of a raid array: level=raid0 devices=0 ctime=Thu Jan 1 07:00:00 1970
mdadm: partition table exists on /dev/sdd but will be lost or meaningless after creating array
Continue creating array? yes
mdadm: Defaulting to version 1.2 metadata
mdadm: array /dev/md0 started.

Instead of making the new array, it starts my array!

➜ ~ ls -l /dev/mapper/MisakaMirror-alldata
lrwxrwxrwx 1 root root 7 май 19 01:48 /dev/mapper/MisakaMirror-alldata -> ../dm-2

But i want to achieve this with

mdadm --assemble

To do it automatically. Thanks.

1 Answer

Resolution

𝕎𝔸ℝℕ𝕀ℕ𝔾: The instructions below delete your existing RAID setup and create a new md RAID 1 array with two entire block devices, /dev/sdc and /dev/sdd.

1. Ensure that your kernel has the RAID 1 md module loaded with either of these commands:

   lsmod | grep 'raid1\s'
   grep 'Personalities : .*\[raid1\]' /proc/mdstat

2. If you do not get an output above, load the RAID 1 md module:

   sudo modprobe raid1

3. Make md forget your existing corrupt array by zapping the disks:

   sudo sgdisk -Z /dev/sdc
   sudo sgdisk -Z /dev/sdd

   Note that mdadm --zero-superblock /dev/sd{c,d} might not operate if mdadm doesn't detect an existing array properly.

4. Recreate your RAID 1 array using the entire devices /dev/sdc and /dev/sdd:

   sudo mdadm --create /dev/md0 -n 2 -l 1 /dev/sdc /dev/sdd

5. To make the new array assemble automatically, add the contents of the following command to the bottom of your /etc/mdadm/mdadm.conf file:

   sudo mdadm --detail --scan

   ^{_{See also: How can I make mdadm auto-assemble RAID after each boot?}}

Explanation

Your two mdadm --misc -E commands reveal that mdadm is not seeing the metadata for your RAID devices. Your example:

➜ ~ sudo mdadm --misc -E /dev/sdc
/dev/sdc: MBR Magic : aa55
Partition[0] : 3907029167 sectors at 1 (type ee)

It looks like /dev/sdc has a partition /dev/sdc1. If you're using a whole device as the md RAID device, you would not have /dev/sdc1. (The same goes for your /dev/sdd.)

Furthermore, when you try recreating the array, mdadm detects this strange information:

mdadm: /dev/sdc appears to be part of a raid array: level=raid0 devices=0 ctime=Thu Jan 1 07:00:00 1970

You're trying to use RAID 1 with two devices today, but mdadm reports RAID 0 with no devices at epoch 0. This is clearly not right.

Maybe at some point you tried creating an array on /dev/sdc1 and /dev/sdd1 (the partitions) rather than on /dev/sdc and /dev/sdd (the entire devices), and the md superblocks got muddled while you struggled to figure out the issue.

For this reason, I believe that you should zap the disks and start over.