RAID controller monitoring - the storage layer below

A RAID controller sits between the OS and the disks. Its health is independent of disk health. This lesson covers how to monitor it.

RAID controller health

A RAID controller has:

• The controller itself (CPU, memory, firmware).
• Cache (RAM, often battery-backed).
• Battery / supercapacitor (for cache persistence).
• Disks (monitored by SMART).
• BBU (battery backup unit) or supercapacitor.

Each can fail. SMART does not cover the controller or the cache.

Vendor tools

Example (Broadcom):

# Show controller info
sudo storcli show

# Show virtual disks
sudo storcli /c0 show vd

# Show physical disks
sudo storcli /c0 show pd

# Show BBU status
sudo storcli /c0 show bbu

What to monitor

For each controller:

• Controller state: optimal, degraded, failed.
• Virtual disk state: optimal, degraded, rebuilding.
• Physical disk state: online, failed, rebuilding.
• BBU state: healthy, charging, failed, missing.
• Cache policy: write-back (with BBU) or write-through.

Alert

# Check RAID state
sudo storcli /c0 show vd | grep -E 'State|rebuilding'
sudo storcli /c0 show bbu | grep -E 'State|charging'

# Alert on:
# - Degraded virtual disk
# - Failed physical disk
# - BBU not healthy
# - Rebuilding state (in progress)

A degraded RAID needs immediate attention. Replace the failed disk and let the rebuild complete.

Cache and BBU

Cache write-back with BBU is the fast path: writes go to RAM, flushed to disk later. Without a healthy BBU, the controller falls back to write-through: every write goes directly to disk, slower.

If the BBU fails:

• Performance drops.
• Risk of data loss on power failure.

Replace the BBU before it fails. Most controllers can be monitored via vendor tools.

Knowledge check