A RAID or NAS warning that the array is “degraded” is easy to underestimate. The files are usually still there, everything still works, and the obvious fix is to put a new drive in and let it rebuild. Most of the time that is fine. But a large share of the RAID cases we see in the lab started as a degraded array and turned into a failed one during the rebuild. Here is how to give yourself the best chance.
What “degraded” means
RAID 1, 5, 6 and 10 store extra information so the array survives a drive failure. Degraded means one (or, in RAID 6, up to two) of the drives has dropped out, and the array is now running without that safety net. RAID 5 that has lost one drive has no remaining redundancy: if another drive fails, or even has a patch of unreadable sectors, the array can go offline.
Why rebuilds fail
Drives in an array are usually the same model, bought on the same day, and have done the same amount of work. When one fails, the others are often not far behind. A rebuild then reads every sector of every remaining drive, for hours or days, under heavy load. That is exactly when a second weak drive tends to give up.
Do this first
- Back up the most important data now, while the array is still readable. Copy the files that matter most first, not everything at once.
- Check the health of the remaining drives in the NAS or controller software. If any show reallocated or pending sectors, or SMART warnings, treat the array as being at high risk.
- Note which bay failed and the order of all drives before you remove anything.
- Only then replace the failed drive and start the rebuild.
Things not to do
- Don’t pull out the wrong drive. Removing a healthy drive from a degraded RAID 5 takes it offline instantly. Double-check bay numbers.
- Don’t re-create or re-initialise the array if it goes offline. This writes new configuration over the old and can make recovery much harder.
- Don’t force a drive that has dropped out back into the array unless you are sure it has the latest data; stale data can corrupt the volume.
- Don’t keep restarting the rebuild if it stops partway through. Stop and get advice.
If the array has already failed
Power it down and label every drive with its bay position. In the lab, each drive is imaged individually – including the failed ones, which may need mechanical work – and the array is rebuilt virtually from those images. The parameters (RAID level, stripe size, drive order, parity rotation) are worked out from the data itself, so we don’t need the original controller or NAS to recover the files.
See our RAID data recovery, NAS data recovery and server data recovery pages, or call 020 7394 2525 before you start a rebuild you’re not sure about.