Wiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== NAS - Build a Linux NAS - Prepare the RAID Disks ======

===== Determine the available disks =====

<code bash>
lsblk
</code>

returns:

<code bash>
NAME                   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda                      8:0    0 232.9G  0 disk 
├─sda1                   8:1    0  1007K  0 part 
├─sda2                   8:2    0   512M  0 part /boot/efi
├─sda3                 253:28   0     8G  0 part [SWAP]
├─sda4                 253:29   0    58G  0 part /
sdb                      8:32   0  14.6T  0 disk 
sdc                      8:48   0  14.6T  0 disk 
sdd                      8:64   0  14.6T  0 disk 
</code>

<WRAP info>
**NOTE:**  This shows:

  * **sda**:  The boot disk of the host OS. Leave this alone!
  * **sdb**:  A free disk.
  * **sdc**:  A free disk.
  * **sdd**:  A free disk.

The free disks do not have any partitions yet and will be included into a RAID.

</WRAP>

----

===== Obtain the Serial Numbers of the Disks =====

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

returns:

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

<WRAP important>
**IMPORTANT:**  Obtain the serial number for every disk; and keep a record of this somewhere safe.
</WRAP>


<WRAP info>
**NOTE:**  Many disks also have the serial number written on the disk.

The reason for recording the serial number of each disk is that in the case of a disk failure, there needs to be a way to determine which disk has failed and needs to be replaced.

  * The enclosure being used for the disks may not support hot-swapping and there may not be a separate light for each disk, so there may not be a straight-forward way to determine which specific disk has failed.

</WRAP>

----

===== Test the Disks =====

All hard drives, both new and old, should be tested before adding them to an array.

See S.M.A.R.T. to view smart data and run smart tests.

Do a short smart test:

<code bash>
smartctl -t short /dev/sdb [-d sat]
</code>

Do a long smart test:

<code bash>
smartctl -t long /dev/sdb [-d sat]
</code>

Check all smart attributes:

<code bash>
smartctl -a /dev/sdb [-d sat]
</code>

Do a random readwrite test:

<code bash>
sudo fio --filename="/dev/sdb --name=randwrite --ioengine=sync --iodepth=1 --rw=randrw --rwmixread=50 --rwmixwrite=50 --bs=4k --direct=0 --numjobs=8 --size=300G --runtime=7200 --group_reporting
</code>

<WRAP info>
**NOTE:** The smartmontools package might be needed to be installed:

<code bash>
sudo apt install smartmontools
</code>

  * **-d sat**:  Sata disks.

See:  https://www.smartmontools.org/

</WRAP>

----

===== Initialize the free disks =====

<code bash>
sudo fdisk /dev/sdb
</code>

returns:

<code bash>
Welcome to fdisk (util-linux 2.36.1).
Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.

Device does not contain a recognized partition table.
The size of this disk is 14.6 TiB (16000900661248 bytes). DOS partition table format cannot be used on drives for volumes larger than 2199023255040 bytes for 512-byte sectors. Use GUID partition table format (GPT).

Created a new DOS disklabel with disk identifier 0xc778227a.

Command (m for help): 
</code>

----

==== Create a GPT Partition ====

<code bash>
Command (m for help): g
</code>

returns:

<code bash>
Created a new GPT disklabel (GUID: 6D811672-A5FE-BA4F-8F79-D17E0285C5E1).
</code>

<WRAP info>
**NOTE:**  GPT (GUID Partition Table) is much better than MBR (Master Boot Record) partitions:

  * GPT supports much larger disks.
  * GPT allows for a nearly unlimited number of partitions. 
  * GPT also stores cyclic redundancy check (CRC) values to check that its data is intact.
    * If the data is corrupted, GPT can notice the problem and attempt to recover the damaged data from another location on the disk.

</WRAP>

----

==== Create a Linux RAID Partition ====

<code bash>
Command (m for help): n
Partition number (1-128, default 1):     
First sector (2048-31251759070, default 2048): 
Last sector, +/-sectors or +/-size{K,M,G,T,P} (2048-31251759070, default 31251759070): 
</code>

returns:

<code bash>
Created a new partition 1 of type 'Linux filesystem' and of size 14.6 TiB.
</code>

<WRAP info>
**NOTE:** Just taking the default values uses the entire disk.
</WRAP>

----

==== Print the Partition ====

<code bash>
Command (m for help): p
</code>

returns:

<code bash>
Disk /dev/sdb: 14.55 TiB, 16000900661248 bytes, 31251759104 sectors
Disk model: ST16000NM001G-2K
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: gpt
Disk identifier: 6D811672-A5FE-BA4F-8F79-D17E0285C5E1

Device     Start         End     Sectors  Size Type
/dev/sdb1   2048 31251759070 31251757023 14.6T Linux filesystem
</code>

<WRAP info>
**NOTE:** This shows that there is a single partition, /dev/sdb1, using the entire disk.
</WRAP>

----

==== Change the Partition Type ====

<code bash>
Command (m for help): t
Selected partition 1
Partition type or alias (type L to list all): 29
</code>

returns:

<code bash>
Changed type of partition 'Linux filesystem' to 'Linux RAID'.
</code>

<WRAP info>
**NOTE:** The type is changed to Linux RAID, which is type 29.


</WRAP>


----

==== Print the Partition again ====

<code bash>
Command (m for help): p
</code>

returns:

<code bash>
Disk /dev/sdd: 14.55 TiB, 16000900661248 bytes, 31251759104 sectors
Disk model: ST16000NM001G-2K
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: gpt
Disk identifier: EAA2F832-9810-AA45-9DDB-8ED531C20139

Device     Start         End     Sectors  Size Type
/dev/sdd1   2048 31251759070 31251757023 14.6T Linux RAID
</code>

<WRAP info>
**NOTE:** This shows the Type has been changed to **Linux RAID**.
</WRAP>

----

==== Write the Partition Table ====

<code bash>
Command (m for help): w
</code>

returns:

<code bash>
The partition table has been altered.
Calling ioctl() to re-read partition table.
Syncing disks.</code>

----

===== Partition other Disks =====

<WRAP info>
**NOTE:** Repeat the above instructions for the other free disks.
</WRAP>


----
===== Check the available disks again =====

<code bash>
lsblk
</code>

returns:

<code bash>
NAME                   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda                      8:0    0 232.9G  0 disk 
├─sda1                   8:1    0  1007K  0 part 
├─sda2                   8:2    0   512M  0 part /boot/efi
├─sda3                 253:28   0     8G  0 part [SWAP]
├─sda4                 253:29   0    58G  0 part /
sdb                      8:32   0  14.6T  0 disk 
├─sdb1                   8:33   0  14.6T  0 part 
sdc                      8:48   0  14.6T  0 disk 
├─sdc1                   8:49   0  14.6T  0 part 
sdd                      8:64   0  14.6T  0 disk 
├─sdd1                   8:65   0  14.6T  0 part 
</code>

<WRAP info>
**NOTE:**  This shows the original free disks all now have partitions.

  * At this point, these disks can be tied together as a RAID array.
</WRAP>