RAID Rebuild Started: Should It Continue?
A healthy rebuild may be allowed to finish. A rebuild that stalls, produces read errors, drops another member, or makes the volume unavailable should not be restarted blindly. The decision depends on what the controller is reading and writing now—not on a fixed percentage.
If the rebuild is stalled, producing read errors, dropping another member, or making the volume unavailable, stop and evaluate before restarting it.
Tell an ADR technician what the system reports and what has already been attempted. The failed system does not have to be brought online first.
What this condition means
The controller reads the surviving members and writes reconstructed blocks to the target. It does not create a protected copy first, and it may not distinguish an isolated bad sector from broader source instability.
A healthy, progressing rebuild can be appropriate. The warning signs are changing member status, repeated retries, abnormal noise, increasing errors, lost access, or a rebuild that repeatedly stops at the same point.
What to record
- Which member failed and which drive is the rebuild target
- Start time, current percentage, and rate of progress
- Controller event log and read-error counts
- Any new degraded, foreign, or missing status
- Whether applications or databases remain active on the volume
What not to do
- Power-cycle solely to restart progress
- Remove the target or another member without recording state
- Begin a second rebuild on a different target
- Run consistency checks concurrently
- Continue production writes when the array is unstable
How ADR evaluates it
ADR reviews the progress, errors, member status, and event sequence to determine whether the rebuild looks healthy or is exposing another failure.
If the rebuild has already changed the array, both the original members and target may contribute to recovery. They can be protected and evaluated separately without requiring the failed array to come online.
ADR can review progress, errors, member health, and the event sequence before the controller is asked to continue. If the rebuild is exposing another failure, the original members and target can be preserved and evaluated separately.
What happens next
If progress and member status are abnormal, ADR can review the evidence before another controller action changes the available data. Start with the current screen, drive list, rebuild percentage, errors, and what happened immediately before the rebuild began.