LSI MegaRAID Rebuild Will Not Start or Fails
A MegaRAID rebuild may refuse to start, stall, or fail when the controller cannot rely on the selected source members and target. The cause can be a second weak member, unreadable sectors, an unsuitable replacement, enclosure or path errors, or metadata that no longer describes a consistent array.
Do not keep changing member states or restarting the rebuild. Preserve the original failed member, the replacement, and the controller event history.
What this condition means
A drive shown as online can still time out or return unreadable sectors under rebuild load. A drive shown as failed may remain partially readable and useful to recovery.
If a rebuild already wrote to the replacement before stopping, that drive contains a partial reconstruction. It should be treated as evidence, not as an empty target.
What to record
- Controller model, firmware, and enclosure
- Virtual-drive and physical-drive status for every member
- Rebuild start, stop percentage, and event-log messages
- Original failed-drive and replacement serial numbers
- Any force-online, foreign import, consistency check, or replacement already attempted
What not to do
- Restart the same rebuild repeatedly
- Discard the original member or partial rebuild target
- Force different members online to test combinations
- Run consistency or filesystem repair
- Initialize or recreate the virtual drive
How ADR evaluates it
ADR identifies which members are stable enough to read and which array generation each may represent. Unstable members are protected before reconstruction work.
The original and partial-rebuild states can then be compared away from the controller, allowing the logical result to be validated without another write to the source set.
What happens next
A failed MegaRAID rebuild does not settle recoverability. The answer depends on readable coverage across the members and how much the attempted rebuild changed.