How to Create a RAID Array with mdadm on a Vultr Cloud Server

Introduction

A Redundant Array of Independent Disks (RAID) is a data technology that combines multiple disk drives to create a single logical storage unit. You can configure a RAID array to improve performance, duplicate data, or both. Usually, the RAID level determines the redundancy or performance improvement and depends on the number of disk drives, configuration, and data distribution across the storage devices.

To increase capacity and performance, create a level, such as RAID 0, that strips and distributes data across multiple disk drives. If your objective is to prevent data loss in case of a storage device failure, create a level that duplicates data across multiple physical disk drives. Typical arrays that prevent data loss include RAID 1, RAID 5, RAID 6, and RAID 10.

A combination of RAID levels, such as RAID 1+0 (RAID 10), provides both performance improvement and data redundancy. The array combines the data stripping of RAID 0 and the mirroring of RAID 1.

You can create almost any RAID level on Ubuntu and other Linux distributions if you have the minimum number of disk drives. This article uses Ubuntu 22.04 LTS to explain how to use the mdadm utility to create various software RAID levels.

Prerequisites

To follow this guide, you need:

1. An Ubuntu 22.04 server that supports multiple storage devices. This could be a Vultr Bare Metal server or a cloud server with multiple block storage volumes attached.
2. A sudo user on the Ubuntu server.
3. Multiple physical hard disks, preferably with the same or almost equal capacities and speed. If you are using a Vultr virtual machine, you can use block storage volumes that operate the same way as physical devices. The minimum number of disk drives depends on the RAID array level. For example, RAID 0 and RAID 1 require a minimum of two drives, while RAID 6 and RAID 10 require at least four devices.

Because we cannot cover all the RAID arrays in one article, this guide focuses on RAID levels 0, 1, 5, and 10.

1: Log in to Server

Connect to your server with SSH as a sudo user.

ssh username@host

If working with previously partitioned disk drives, you need to delete existing partitions and create those that match your intended array size and configuration.

2: Check if RAID mounted

Check if there are previous RAID arrays and unmount them. If using new disk drives, skip this and go to step 3.

$ sudo cat /proc/mdstat   

Output

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]  
   unused devices: <none>  

If there is an existing RAID array. Unmount it using the syntax;

$ sudo unmount /dev/mdx

For example, for a RAID name md0, use the command:

$ sudo umount /dev/md0

To stop the RAID array md0, run:

$ sudo mdadm --stop /dev/md0

Remove the array so that you can access and reuse the disk drives.

$ sudo mdadm --remove /dev/md0 

3: Identify available disk drives

Run the following to identify the available disk drives, sizes, file systems, and mountpoints.

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT  

Typical output

NAME    SIZE FSTYPE TYPE MOUNTPOINT  
sr0    1024M        rom  
vda      50G        disk  
└─vda1   50G ext4   part /  
vdb     160G        disk  
vdc     160G        disk 

In addition to the drive vda with the operating system, we have two more devices, vdb and vdc.

4: Partition the disk drives

Before creating the RAID, it is good practice to partition disk drives such as vdb and vdc.

Create an empty partition table on vdb.

$ sudo parted -s /dev/vdb mklabel gpt  

Run the same command for the second disk drive, vdc.

$ sudo parted -s /dev/vdc mklabel gpt

If you have more than two drives, use the above command for all of them while replacing vdc with the appropriate disk identifier.

Create the maximum partition on the drive. However, you can specify a smaller partition if you do not want to use the entire disk drive.

$ sudo parted -s /dev/vdb unit mib mkpart primary 0% -100%

Repeat the command for all the drives by replacing vdb with the appropriate disk drive name.

5: Create software RAID

The syntax for creating any RAID level is:

$ sudo mdadm --create --verbose /dev/[ RAID array Name] --level=[RAID Level] --raid-devices=[Number of storage devices] [Storage Device Identifier] [Storage Device Identifier]

The verbose flag, which displays the progress in real-time, is optional.

5.1: How to create RAID 0

RAID 0 requires a minimum of two storage devices.

Scan for available disk drives, partitions, and their identifiers.

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT

Output

     NAME    SIZE FSTYPE TYPE MOUNTPOINT  
sr0    1024M        rom 
vda      50G        disk 
    └─vda1   50G ext4   part / 
vdb     160G        disk  
└─vdb1  160G        part
vdc     160G        disk  
└─vdc1  160G        part 

To create RAID 0 called md0 using the primary partitions vdb1 and vdc1, run the command;

$ sudo mdadm --create --verbose /dev/md0 --level=0 --raid-devices=2 /dev/vdb1 /dev/vdc1  

Typical output

mdadm: chunk size defaults to 512K  
mdadm: Defaulting to version 1.2 metadata  
mdadm: array /dev/md0 started.  

Check RAID status.

$ sudo cat /proc/mdstat  

Output

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]  
md0 : active raid0 vdc1[1] vdb1[0] 
      335276032 blocks super 1.2 512k chunks 

unused devices: <none> 

The output shows the active RAID level and disk drives/partition it uses.

To configure automatic assembly and mounting of the RAID 0 array on boot, go to step 6.

5.2: How to create RAID 1

RAID 1 requires a minimum of two storage devices.

Scan for available disk drives, partitions, and their identifiers.

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT   

Output

NAME    SIZE FSTYPE TYPE MOUNTPOINT  
sr0    1024M        rom  
vda      25G        disk  
└─vda1 25G ext4 part /
vdb 120G disk
vdc 120G disk

Although you can use partitions, such as in RAID 0 above, this section shows how to create RAID 1 using 120 GB disk drives without partitioning them first.

 $ sudo mdadm --create --verbose /dev/md0 --level=1 --raid-devices=2 /dev/vdb /dev/vdc 

Type `y' for yes when prompted.

Output

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  
    our boot-loader understands md/v1.x metadata, or use  
    --metadata=0.90  
mdadm: size set to 125762560K  
mdadm: automatically enabling write-intent bitmap on large array  
Continue creating array? y
mdadm: Defaulting to version 1.2 metadata  
mdadm: array /dev/md0 started.  

Check the RAID status.

$ sudo cat /proc/mdstat  

Typical output:

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]  
md0 : active raid1 vdc[1] vdb[0]  
      125762560 blocks super 1.2 [2/2] [UU]  
      [=>...................]  resync =  9.2% (11666368/125762560) finish=19.5min speed=97113K/sec  
      bitmap: 1/1 pages [4KB], 65536KB chunk  

unused devices: <none>  

To display the list of disk drives, size assigned to RAID, and status, run:

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT 

Typical output

NAME     SIZE FSTYPE            TYPE  MOUNTPOINT  
sr0     1024M                   rom  
vda       25G                   disk  
└─vda1    25G ext4              part  /  
vdb      120G linux_raid_member disk  
└─md0  119.9G                   raid1  
vdc      120G linux_raid_member disk  
└─md0  119.9G                   raid1  

To configure automatic assembly and mounting of the RAID 1 array on boot, go to step 6.

5.3: How to create RAID 5

RAID 5 requires a minimum of three storage devices.

Scan for available disk drives, partitions, and their identifiers.

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT 

Typical output

NAME    SIZE FSTYPE TYPE MOUNTPOINT  
sr0    1024M        rom  
vda      50G        disk  
└─vda1   50G ext4   part /  
    vdb     160G        disk  
vdc     160G        disk  
vdd     160G        disk

We assume they are all new and thus create RAID 5 without prior partitioning.

$ sudo mdadm --create --verbose /dev/md0 --level=5 --raid-devices=3 /dev/vdb /dev/vdc /dev/vdd   

Check active RAID

$ sudo cat /proc/mdstat

Output

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]  
md0 : active raid5 vdd[3] vdc[1] vdb[0]  
      335280128 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/2] [UU_]  
      [>....................]  recovery =  3.4% (5773616/167640064) finish=31.3min speed=86110K/sec  
      bitmap: 0/2 pages [0KB], 65536KB chunk  

To configure automatic assembly and mounting of the RAID 5 array on boot, go to step 6.

5.4: How to create RAID 10

RAID 10 requires a minimum of four storage devices.

Scan for available disk drives, partitions, and their identifiers.

$ lsblk -o NAME,SIZE,FSTYPE,TYPE,MOUNTPOINT

Output

        NAME    SIZE FSTYPE TYPE MOUNTPOINT  
 sr0    1024M        rom  
 vda      50G        disk  
     └─vda1   50G ext4   part /
vdb     160G        disk 
vdc     160G        disk
vdd     160G        disk
vde     160G        disk

Create the RAID 10 array

$ sudo mdadm --create --verbose /dev/md0 --level=10 --raid-devices=4 /dev/vdb /dev/vdc /dev/vdd /dev/vde

Check active RAID

$ sudo cat /proc/mdstat

Output

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]  
md0 : active raid10 vde[3] vdd[2] vdc[1] vdb[0]  
      335280128 blocks super 1.2 512K chunks 2 near-copies [4/4] [UUUU]  
      [>....................]  resync =  2.6% (8807232/335280128) finish=32.4min speed=167884K/sec  
      bitmap: 3/3 pages [12KB], 65536KB chunk  

6: Create and mount a file system on the RAID array

Once you configure a RAID array, you must create a file system and mount the array. The following instructions show you how to create the file system, mount the RAID array, and configure it to assemble and mount the RAID array when booting. The commands are the same for all levels.

Create the file system

The syntax for creating the file system is;

$ sudo mkfs –t [File system type] [RAID Device]

For example, to create the Ext4 file system on RAID 1 we created above, run the command;

$ sudo mkfs -t ext4 /dev/md0 

7: Mount the RAID array

To mount the RAID array, run:

$ sudo mount /dev/md0 /mnt/

Check whether the new file system is accessible.

$ df -h -x devtmpfs -x tmpfs

8: Configure automatic mount on boot

To ensure that the server automatically assembles and mounts the RAID array when booting, you need to edit the configuration file /etc/mdadm/mdadm.conf.

Use the command below to scan the system, identify the active array and automatically make the necessary changes to the `/etc/mdadm/mdadm.conf file.

$ sudo mdadm --detail --scan | sudo tee -a /etc/mdadm/mdadm.conf

If you want the array loads early in the boot, update the initial RAM file system (initramfs).

$ sudo update-initramfs -u

Additionally, you must add the file system mount options to the etc/fstab file. Run the command below to get the array details and automatically add them to the etc/fstab file.

$ echo '/dev/md0 /mnt/md0 ext4 defaults,nofail,discard 0 0' | sudo tee -a /etc/fstab

The above commands ensure automounting of the RAID array when the machine boots and applies to all the RAID levels. To confirm, reboot the server and run the command below to get the RAID status.

$ sudo cat /proc/mdstat

Summary

A RAID array is a storage technology that combines multiple physical disk drives to create a logical unit that offers data redundancy and recovery, performance improvement, or both. The choice of the RAID array level depends on whether you want to improve the read/write performance, add redundancy and prevent data loss in case of a drive failure. While the article focuses on a few RAID levels, you can use the instructions to create other arrays in Ubuntu 22.04 LTS and other Linux distributions.