====== NAS - Build a Linux NAS - Support recovering from a faulty disk ======

===== Determine which disk has failed =====

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

returns:

<code bash>
</code>

----

===== Get the serial number of the failed disk =====

<code bash>
hdparm -i /dev/sdd | grep SerialNo
</code>

returns:

<code bash>
Model=ST16000NM001G-2KK103, FwRev=SB30, SerialNo=WL2091XL
</code>

<WRAP info>
**NOTE:**  Double check which actual physical disk has failed and needs to be replaced.
</WRAP>

----

===== Replace the faulty disk =====

Steps are:

  * Wait until the disks has finished rebuilding on the spare disk if any.
  * Power the server down.
  * Replace the faulty disk.
  * Power the server up again.
  * Partition the disks.
  * Add the new disk to the array with: <code bash>
mdadm --add /dev/md0 /dev/sdd1
</code>

<WRAP info>
**NOTE:**  To add a failed drive, for example /dev/sdd1 back to the RAID the following can also be used:

<code bash>
sudo mdadm --manage /dev/md0 --add /dev/sdd1
</code>

</WRAP>


----

===== Recover the Raid =====

If the original OS dies, but the RAID disks are okay, then rebuild a new OS.

Then, use mdadm to examine the superblock on one of the disks and use this to recover the raid:

<code bash>
mdadm --examine /dev/sda
</code>

returns:

<code bash>
</code>

Then use the listed UUID from the disk to assemble the raid onto the device md0:

<code bash>
mdadm --assemble /dev/md0 --uuid=<UUID>
</code>

<WRAP info>
**NOTE:**  See [[https://www.linuxquestions.org/questions/linux-general-1/recovering-mdadm-superblocks-713234/|Recovering mdadm superblocks]]
</WRAP>