RAID 10 Offline After Power Loss: What to Preserve First
A power event can leave a RAID 10 controller unable to present the virtual disk even when the members still contain the data. The outage may involve controller cache, configuration metadata, enclosure communication, or disagreement inside one or more mirror pairs.
Do not recreate or initialize the array. Record the controller view, photograph the bays, and keep every member in its original position.
What this condition means
RAID 10 needs one usable, current member from each mirror pair and the correct stripe relationship between the pairs. A simple count of healthy drives does not prove that those requirements are met.
After power loss, members may appear foreign, missing, unconfigured, or healthy while carrying different write histories. Controller labels are evidence to record, not instructions to accept automatically.
What to record
- Server, enclosure, controller, firmware, and RAID 10 member count
- Virtual-disk and physical-drive status screens
- Slot and serial number of every drive
- Cache, battery, capacitor, and power-event messages
- Any reboot, import, clear, force-online, recreation, or repair already attempted
What not to do
- Create a replacement RAID 10 with remembered settings
- Initialize drives shown as new or unconfigured
- Import or clear foreign metadata without reviewing the complete set
- Move members between bays while testing
- Run filesystem repair on a volume whose pair relationships are unverified
How ADR evaluates it
ADR compares controller evidence with member metadata, readability, mirror-pair relationships, and write generation.
The server does not need to boot and the failed array does not need to be brought online. A validated RAID 10 reconstruction can be built from protected member access and recovered to separate storage.
What happens next
Power loss often changes how the controller presents a RAID 10 rather than erasing every member. Preserving the complete set keeps configuration and reconstruction options available.