mdadm is the standard Linux utility for managing software RAID arrays. It creates, monitors, and manages MD (multiple device) arrays without needing a dedicated hardware RAID controller – making it a cost-effective way to add disk redundancy and performance to any server.
This guide covers installing mdadm, creating RAID 1, RAID 5, and RAID 10 arrays, formatting and mounting, persisting the configuration, setting up email monitoring alerts, and replacing failed disks on Rocky Linux 10 and AlmaLinux 10. The same steps apply to RHEL 10.
Prerequisites
- A server running Rocky Linux 10, AlmaLinux 10, or RHEL 10
- Root or sudo access
- At least 2 unused disks for RAID 1, 3 for RAID 5, or 4 for RAID 10
- A working mail relay if you want email alerts (postfix, mailx, or similar)
RAID Level Quick Reference
| RAID Level | Method | Min Disks |
|---|---|---|
| RAID 1 (Mirror) | Mirroring – identical copy on each disk | 2 |
| RAID 5 (Parity) | Striping with distributed parity | 3 |
| RAID 10 (1+0) | Mirrored pairs then striped | 4 |
Step 1: Install mdadm on Rocky Linux 10 / AlmaLinux 10
Install the mdadm package and confirm it is available.
sudo dnf install mdadm -yVerify the installation.
mdadm --versionStep 2: Identify Available Disks
List all block devices to find the disks you will use for the RAID array.
lsblkSample output showing four 10 GB disks ready for RAID.
$ lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
sda 8:0 0 40G 0 disk
├─sda1 8:1 0 1G 0 part /boot
└─sda2 8:2 0 39G 0 part
├─rl-root 253:0 0 35G 0 lvm /
└─rl-swap 253:1 0 4G 0 lvm [SWAP]
sdb 8:16 0 10G 0 disk
sdc 8:32 0 10G 0 disk
sdd 8:48 0 10G 0 disk
sde 8:64 0 10G 0 diskIn this example we have /dev/sdb, /dev/sdc, /dev/sdd, and /dev/sde. Replace these with your actual device names throughout the guide.
Step 3: Prepare Disks with Partition Tables
Create GPT partition tables and a single partition on each disk. Install parted if it is not already present.
sudo dnf install parted -ySwitch to root to run partition commands without repeating sudo.
sudo -iCreate a GPT label, a full-disk partition, and set the RAID flag on each disk.
for disk in /dev/sdb /dev/sdc /dev/sdd /dev/sde; do
parted --script "$disk" "mklabel gpt"
parted --script "$disk" "mkpart primary 0% 100%"
parted --script "$disk" "set 1 raid on"
doneVerify the partition layout on one of the disks.
parted --script /dev/sdb "print"Expected output.
$ parted --script /dev/sdb "print"
Model: ATA VBOX HARDDISK (scsi)
Disk /dev/sdb: 10.7GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Disk Flags:
Number Start End Size File system Name Flags
1 1049kB 10.7GB 10.7GB primary raidStep 4: Create a Software RAID Array with mdadm
Choose one of the RAID levels below depending on your requirements. Each example creates an array at /dev/md0.
Create RAID 1 (Mirror – 2 disks)
RAID 1 mirrors data across two disks. If one disk fails, the other has a full copy. Usable capacity is half the total disk space.
mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/sdb1 /dev/sdc1Create RAID 5 (Striping with Parity – 3 disks)
RAID 5 stripes data with distributed parity across three or more disks. One disk can fail without data loss. Usable capacity is (N-1) disks.
mdadm --create /dev/md0 --level=5 --raid-devices=3 /dev/sdb1 /dev/sdc1 /dev/sdd1Create RAID 10 (Mirror + Stripe – 4 disks)
RAID 10 combines mirroring and striping. It needs at least four disks and can tolerate one disk failure per mirrored pair. Usable capacity is half the total. This is the recommended level for database servers and high-I/O workloads.
mdadm --create /dev/md0 --level=10 --raid-devices=4 /dev/sdb1 /dev/sdc1 /dev/sdd1 /dev/sde1When prompted with “Continue creating array?”, type y and press Enter. You will see output confirming the array started.
mdadm: Defaulting to version 1.2 metadata
mdadm: array /dev/md0 started.Step 5: Verify the RAID Array
Check the array status with mdadm --detail. This is the primary command for inspecting any MD array. If you manage LVM logical volumes on top of RAID, the detail output helps confirm the underlying array health.
mdadm --detail /dev/md0Sample output for a RAID 10 array.
$ mdadm --detail /dev/md0
/dev/md0:
Version : 1.2
Creation Time : Sat Mar 21 10:15:32 2026
Raid Level : raid10
Array Size : 20953088 (19.98 GiB 21.46 GB)
Used Dev Size : 10476544 (9.99 GiB 10.73 GB)
Raid Devices : 4
Total Devices : 4
Persistence : Superblock is persistent
Update Time : Sat Mar 21 10:17:00 2026
State : clean
Active Devices : 4
Working Devices : 4
Failed Devices : 0
Spare Devices : 0
Layout : near=2
Chunk Size : 512K
Consistency Policy : resync
Name : server01:0 (local to host server01)
UUID : 089d8e15:403e261e:dbbee899:36a357b2
Events : 17
Number Major Minor RaidDevice State
0 8 17 0 active sync set-A /dev/sdb1
1 8 33 1 active sync set-B /dev/sdc1
2 8 49 2 active sync set-A /dev/sdd1
3 8 65 3 active sync set-B /dev/sde1You can also check the kernel’s view of all MD arrays.
cat /proc/mdstatExamine individual member disks with the --examine flag.
mdadm --examine /dev/sdb1 /dev/sdc1 /dev/sdd1 /dev/sde1Step 6: Create a Filesystem and Mount the RAID
Format the array with ext4 (or xfs if you prefer the default Rocky Linux filesystem).
mkfs.ext4 /dev/md0Create a mount point and mount the array.
mkdir -p /mnt/raid
mount /dev/md0 /mnt/raidVerify the mount.
$ df -hT /mnt/raid
Filesystem Type Size Used Avail Use% Mounted on
/dev/md0 ext4 20G 45M 19G 1% /mnt/raidStep 7: Save mdadm Configuration
Write the array definition to /etc/mdadm.conf so the system reassembles the array automatically on boot.
mdadm --detail --scan >> /etc/mdadm.confVerify the content was written.
cat /etc/mdadm.confYou should see a line like this.
ARRAY /dev/md0 metadata=1.2 name=server01:0 UUID=089d8e15:403e261e:dbbee899:36a357b2Step 8: Add Persistent Mount in /etc/fstab
Add the RAID device to /etc/fstab for automatic mounting at boot. Using the UUID is more reliable than the device name.
blkid /dev/md0Copy the UUID from the output, then add a line to fstab. Open the file.
sudo vi /etc/fstabAdd the following line at the end (replace the UUID with your actual value).
UUID=e2570369-67d8-444c-9125-6afdb68dc7d5 /mnt/raid ext4 defaults 0 0Test the fstab entry without rebooting.
umount /mnt/raid
mount -a
df -hT /mnt/raidStep 9: Monitor RAID with Email Alerts
mdadm includes a built-in monitoring daemon that watches for disk failures, spare activations, and rebuild events. When something goes wrong, it sends an email alert so you can respond before data is at risk.
Install mailx for sending alerts
The monitoring daemon needs a mail command to send notifications. Install mailx and make sure your system can relay mail (through Postfix or a similar MTA).
sudo dnf install mailx -yConfigure the mdadm monitor daemon
Open the mdadm configuration file.
sudo vi /etc/mdadm.confAdd a MAILADDR line with your email address at the top of the file.
MAILADDR [email protected]Enable and start the mdadm monitoring service.
sudo systemctl enable mdmonitor --nowVerify the service is running.
sudo systemctl status mdmonitorYou can test the monitoring manually by sending a test alert.
mdadm --monitor --scan --test --oneshotCheck your inbox for the test email. If it does not arrive, verify that your MTA is configured and that port 25 (TCP) is open on the server’s firewall.
Step 10: Simulate and Replace a Failed Disk
Knowing how to replace a failed disk is critical for production RAID arrays. Here is the full procedure.
Mark a disk as failed (simulate failure)
For testing purposes, you can mark a disk as faulty.
mdadm --manage /dev/md0 --fail /dev/sdd1Check the array status to confirm the failure.
$ mdadm --detail /dev/md0
State : clean, degraded
Number Major Minor RaidDevice State
0 8 17 0 active sync set-A /dev/sdb1
1 8 33 1 active sync set-B /dev/sdc1
- 0 0 2 removed
3 8 65 3 active sync set-B /dev/sde1
2 8 49 - faulty /dev/sdd1Remove the failed disk
mdadm --manage /dev/md0 --remove /dev/sdd1Add a replacement disk
After physically replacing the failed drive (or using the same device name for testing), partition the new disk the same way as in Step 3, then add it to the array.
parted --script /dev/sdd "mklabel gpt"
parted --script /dev/sdd "mkpart primary 0% 100%"
parted --script /dev/sdd "set 1 raid on"Add the new partition to the array.
mdadm --manage /dev/md0 --add /dev/sdd1The array will begin rebuilding automatically. Monitor the rebuild progress.
$ cat /proc/mdstat
Personalities : [raid10]
md0 : active raid10 sdd1[4] sde1[3] sdc1[1] sdb1[0]
20953088 blocks super 1.2 512K chunks 2 near-copies [4/3] [UU_U]
[=====>...............] recovery = 28.3% (2965504/10476544) finish=3.2min speed=38945K/secWait for the rebuild to complete, then verify the array is clean again.
mdadm --detail /dev/md0The State should show clean with all devices listed as active sync.
Step 11: Add a Hot Spare Disk
A hot spare sits idle in the array until a disk fails, then automatically takes over the failed disk’s role and starts rebuilding. This reduces the window of vulnerability.
Prepare the spare disk.
parted --script /dev/sdf "mklabel gpt"
parted --script /dev/sdf "mkpart primary 0% 100%"
parted --script /dev/sdf "set 1 raid on"Add it as a spare.
mdadm --manage /dev/md0 --add /dev/sdf1Verify the spare is recognized. It will show as spare in the detail output.
$ mdadm --detail /dev/md0 | tail -8
Number Major Minor RaidDevice State
0 8 17 0 active sync set-A /dev/sdb1
1 8 33 1 active sync set-B /dev/sdc1
4 8 49 2 active sync set-A /dev/sdd1
3 8 65 3 active sync set-B /dev/sde1
5 8 81 - spare /dev/sdf1Useful mdadm Commands Reference
| Command | Purpose |
|---|---|
mdadm --detail /dev/md0 | Full array status and disk list |
cat /proc/mdstat | Kernel-level array status and rebuild progress |
mdadm --examine /dev/sdb1 | Show superblock info on a member disk |
mdadm --manage /dev/md0 --fail /dev/sdb1 | Mark a disk as failed |
mdadm --manage /dev/md0 --remove /dev/sdb1 | Remove a failed disk from the array |
mdadm --manage /dev/md0 --add /dev/sdg1 | Add a new or spare disk to the array |
mdadm --stop /dev/md0 | Stop (deactivate) an array |
mdadm --assemble --scan | Reassemble all arrays from mdadm.conf |
Conclusion
You now have a working software RAID array on Rocky Linux 10 / AlmaLinux 10 managed by mdadm, with persistent mounting, email monitoring alerts, and a clear disk replacement procedure. You can also layer LVM on top of RAID for flexible volume management.
For production servers, consider adding a hot spare to every array, regularly checking /proc/mdstat for degraded arrays, testing your email alerts quarterly, and keeping a documented runbook for disk replacement procedures. Refer to the Linux kernel md documentation for advanced configuration options.
