RAID 5 Rebuild or Image First? The Decision That Matters
Every degraded array asks the same question: rebuild now, or image first? The answer depends on the health of the surviving members, not on how urgent the data feels.
Rebuild is a full-surface read
A rebuild reads every sector of every member. On aging disks this is the highest-stress operation the array will ever see, and weak sectors are exactly where it stalls.
Signals that say 'do not rebuild'
SMART reallocations rising, any member with a history of timeouts, mixed firmware or mixed batches, or an array already on its second failure. In these cases image the members read-only and assemble virtually.
When rebuild is acceptable
Healthy members with clean SMART history, verified backups, and a maintenance window. Even then, capture a metadata dump and note disk order before starting.
FAQ
Can I stop a rebuild midway?
Yes, and sometimes you should — but powering down mid-write leaves the array in an inconsistent state. Decide before starting.
Does a hardware controller make it safer?
It makes it faster, not safer. The read stress is the same.
What if the rebuild already failed?
Stop, image every member, and treat it as a parameter-recovery case.
Boundary: once a rebuild has written inconsistent parity across members, only file-level salvage may remain.
Free evaluation in our Shenzhen lab; written quotation after inspection; no recovery, no fee. Mail-in service available for Hong Kong and overseas clients.