====== RAID - mdadm - Add a disk to an existing array ======

Add a new disk to an array, possibly replacing a failed one:

<code bash>
mdadm --add /dev/md0 /dev/sdb1
</code>

----

===== Add a disk using its UUID =====

==== Check the UUIDs of the partitions ====

<code bash>
ls -l /dev/disk/by-partuuid/
</code>

returns:

<code bash>
lrwxrwxrwx 1 root root 10 Sep  4 14:49 526e2405-0124-4baa-805f-335369588ddd -> ../../sda1
lrwxrwxrwx 1 root root 10 Sep  4 14:49 5d426ea0-5795-4165-863d-1ed947dc77fd -> ../../sda2
lrwxrwxrwx 1 root root 10 Sep  4 14:49 8e53314c-0814-46cd-9f13-c2bb2b7a14b6 -> ../../sdb1
lrwxrwxrwx 1 root root 10 Sep  4 14:49 90412d26-ebec-4996-96d6-a7a6c702e57e -> ../../sdc1
</code>

----

==== Add /dev/sdc1 using its UUID ====

<code bash>
mdadm /dev/md0 --add /dev/disk/by-partuuid/90412d26-ebec-4996-96d6-a7a6c702e57e
</code>

returns:

<code bash>
mdadm: added /dev/disk/by-partuuid/90412d26-ebec-4996-96d6-a7a6c702e57e
</code>

----

==== Check the details of the RAID array ====

<code bash>
mdadm --detail /dev/md0
</code>

returns:

<code bash>
</code>

<WRAP info>
**NOTE:**  It reports /dev/sdc1 instead of the UUID.

This is not a problem.

mdadm allows disks to be moved around freely in the machine, regardless of how you add the disk to the array.

  * It tracks the disks by the RAID metadata (superblocks) stored on the disk.
  * The default config, specified in /etc/mdadm/mdadm.conf, usually is normally DEVICE partitions, which means to look at all partitions (on all disks) checking for RAID superblocks.
    * It checks for a match of the array name or UUID (depending on what is in that config file.
</WRAP>