RAID Configuration: Data Protection on Servers

RAID (Redundant Array of Independent Disks) protects data by distributing it across multiple drives. If one drive fails, your data survives. RAID is essential for dedicated servers where uptime and data protection are critical.

RAID Levels Overview

• RAID 0 (Striping): Speed only, no protection. Fails if ANY drive fails.
• RAID 1 (Mirroring): Complete copy on 2nd drive. 50% usable space. Protected against 1 drive failure.
• RAID 5 (Striping + Parity): Distributed data + parity across 3+ drives. Protected against 1 failure. 67% usable space with 3 drives.
• RAID 6 (Striping + Dual Parity): Like RAID 5 but survives 2 drive failures. Requires 4+ drives. 50% usable space.
• RAID 10 (1+0): Mirrors + striping. Fast, protected against 1 failure per mirror pair. Requires 4+ drives.

Comparison Table

Hardware vs Software RAID

Hardware RAID (Recommended):

• Dedicated RAID controller card
• Works even if OS crashes
• Better performance
• Battery-backed cache (protects during power loss)
• Higher cost, but professional-grade

Software RAID (Linux md):

• No special hardware needed
• CPU-dependent (uses server resources)
• Simpler to rebuild
• Lower cost
• Good for small servers, cost-conscious setups

Configure RAID on Dedicated

Hardware RAID configuration:

Usually configured during server provisioning in BIOS/UEFI:

1. Contact UnderHost support for RAID setup during order
2. Specify RAID level (RAID 5 recommended)
3. OS installs on pre-configured RAID array

Linux Software RAID:

# Create RAID 5 array
sudo mdadm --create /dev/md0 --level=5 --raid-devices=3 /dev/sda /dev/sdb /dev/sdc

# Check status
cat /proc/mdstat
sudo mdadm --detail /dev/md0

# Save configuration
sudo mdadm --detail --scan >> /etc/mdadm.conf

Monitor RAID Health

# Check RAID status
cat /proc/mdstat

# Detailed status
sudo mdadm --detail /dev/md0

# Check for errors
sudo dmesg | grep -i raid

Drive Failure Recovery

RAID 5 survives 1 drive failure. Replace failed drive:

# Identify failed drive
sudo mdadm --detail /dev/md0

# Remove failed drive
sudo mdadm /dev/md0 --fail /dev/sda --remove /dev/sda

# Add replacement
sudo mdadm /dev/md0 --add /dev/sdd

# Monitor rebuild (can take hours)
cat /proc/mdstat

For hardware RAID: Contact UnderHost support to hot-swap the drive.

Best Practices

• RAID ≠ Backup: RAID protects against hardware failure only, not data deletion or corruption
• Monitor constantly: Watch for drive failures
• Keep spare drives: Reduces rebuild time during failure
• Test recovery: Practice drive replacement before production failure
• Combine with backups: RAID + offsite backups = complete protection

Related: Backup strategy | Disaster recovery | Dedicated servers