RAID 5 Rebuild Stuck: When to Stop and Evaluate
A RAID 5 rebuild that stops progressing is often waiting on repeated reads, encountering an unstable member, losing communication with a drive, or failing to write reliably to the target. The percentage alone does not reveal which condition is present.
Do not power-cycle solely to restart the counter or begin another rebuild. Record the percentage, errors, member status, and drive order while the current state is still visible.
What this condition means
Rebuild activity reads every required region from the surviving set. That load can expose unreadable sectors or intermittent behavior that ordinary operation did not reveal.
A stalled target may contain valid reconstructed blocks up to a point and incomplete data beyond it. Both the target and the original member set may matter during recovery.
What to record
- Controller, RAID level, member count, and enclosure
- Rebuild percentage, elapsed time, and whether it still changes
- Slot, serial number, and current status of every drive
- Controller log entries, timeouts, and media errors
- Which member was replaced and whether the volume remains accessible
What not to do
- Restart or repeat the rebuild as a test
- Remove the target without identifying it
- Replace another member while the cause is unknown
- Run a consistency check at the same time
- Continue heavy production writes on the degraded array
How ADR evaluates it
ADR reviews the failure sequence and member health to determine whether the rebuild should continue, be stopped, or give way to image-first recovery.
When the rebuild has already altered the set, the original members and target are evaluated separately and reconstruction is tested against protected sources.
What happens next
A stuck rebuild can still leave a recoverable member set. Preserving the original members and partial target gives the technician more options than another uncontrolled attempt.