RAID Cache Battery Failure: Fix in 10 Minutes

Hardware – RAM & MB Intermediate 👁 29 views 📅 Jun 29, 2026

RAID controller says cache battery is dead. Your RAID write performance drops to crap. Replace or reset it, or force write-back mode.

Quick answer: Replace the battery module or run a learn cycle. If that fails, force write-back mode via the RAID utility (but you risk data loss on power loss).

This happens more often than you'd think. RAID cache batteries are basically rechargeable lithium-ion cells. They die after 2-4 years. When they go, the controller panics and drops to write-through mode. That means every write hits the disk directly. Your RAID 5 array write speed goes from 800MB/s to 40MB/s. The controller logs a "BBU failed" or "Battery backup unit voltage low" warning. Users complain about slow file copies. You check the RAID software and see a red battery icon.

Fix Steps

  1. Identify the controller model. Open the server chassis. Look for a Dell PERC H710, H730, H740, HP Smart Array P420, or LSI MegaRAID 9361. Write down the model number. You need the exact battery part number. For example, Dell PERC H730 uses the 0H4T8 battery. HP P420 uses the 734317-001.
  2. Check the battery voltage. Use the RAID utility. On Dell: omreport storage battery controller=0. On LSI: storcli /c0/bbu show. Look at "Voltage" and "Temperature". If voltage is below 3.0V or temperature > 60C, the battery is dead.
  3. Run a learn cycle (if voltage is borderline). A learn cycle discharges and recharges the battery to recalibrate it. On LSI: storcli /c0/bbu start learn. On Dell: omconfig storage battery action=learn controller=0. Wait 4-6 hours. Check status again. If still failing, replace it.
  4. Replace the battery module. Power down the server. Disconnect power cables. Pop the old battery out. It's usually a small board with a connector. Slot the new one in. Power up. The controller should detect it. Run a learn cycle manually after 24 hours.
  5. Force write-back mode (temporary workaround). If you need performance back NOW and cannot wait for a replacement, force write-back. On LSI: storcli /c0 set wb=on. On Dell: omconfig storage controller action=setwritepolicy controller=0 writepolicy=writeback. Warning: This disables battery protection. If power drops, you lose any data in cache. Only do this on servers with UPS that you trust.

Alternative Fixes If Main One Fails

Bad connector or corrosion

Sometimes the battery is fine but the connector has corrosion. Take the battery out. Clean the pins with isopropyl alcohol and a toothbrush. Reconnect. If the voltage reading jumps up, that was your problem. Happens on older servers in humid environments.

Controller firmware bug

Some older PERC H700 and H710 units had a firmware bug that falsely reported a dead battery. Update firmware to the latest version. Find it on Dell's support site. Flash it via the bootable ISO. I've seen this fix 1 in 20 cases.

Cache module failure (rare)

If replacing the battery doesn't work, the actual cache module on the RAID card might be dead. Replace the whole RAID card. But test the battery first. It's cheaper.

Prevention Tips

  • Schedule learn cycles. Run them every 3 months automatically. Most RAID utilities let you schedule this. LSI: storcli /c0/bbu set autolearnperiod=90. This keeps the battery calibrated.
  • Replace batteries proactively every 3 years. Don't wait for failure. Mark it in your calendar. Batteries are cheap compared to downtime.
  • Monitor battery health. Use SNMP traps or nagios plugins. Alert on voltage below 3.5V or temperature above 50C. I use check_mk with a custom script for this.
  • Keep a spare battery on your shelf. For your most common server model. You'll thank me when it fails at 3 AM and you swap it in 5 minutes instead of waiting for overnight shipping.
  • Document the part numbers. Print them and tape them inside the server chassis. Future you or a junior colleague will appreciate it.
My rule of thumb: If the battery is over 3 years old and showing signs of failure, replace it. Don't bother trying to revive it with learn cycles. Those only work for batteries that are slightly off-calibration, not dead ones.

RAID cache battery failure is a common issue that trashes your write performance. But it's straightforward to fix. Replace the battery, run a learn cycle, or force write-back temporarily. Just don't ignore it — your users will keep complaining about slow file transfers.

Was this solution helpful?