Rebuilds do not fail randomly. They fail because something in the array cannot be read, and running it again applies the same maximum load to the same struggling disks — usually with a worse result the second time.
It sweeps every surviving member end to end and writes as it goes. On disks already carrying unreadable sectors that is simultaneously the likeliest moment for a second failure and the point at which wrong data gets committed.
A rebuild that will not start, stalls partway or fails repeatedly is not usually a controller problem. It is the array reporting that the surviving disks cannot supply clean reads across their entire surface — which is exactly what a rebuild requires.
That is worth understanding before trying again. Each failed attempt is another full-surface pass over disks that have already demonstrated they cannot manage one. Repeated attempts are how a degraded array becomes a dead one.
If a rebuild has failed once, do not start it again. The information you needed has already been supplied.
Take the array out of service and image every member individually. Imaging reads each disk once, gently, working around bad areas rather than hammering them — the opposite of what a rebuild does.
Once every member exists as an image, the array is reconstructed from those copies. Nothing further is risked, failed reads can be retried selectively, and the surviving disks are never asked to do anything again.
Where the business genuinely cannot wait, that is a legitimate reason to attempt another rebuild. But make it a decision rather than a reflex, knowing that a second failure during it is likely rather than possible.
A surviving member has unreadable sectors at that point. Each retry is another full pass over disks already struggling. Stop and image instead.
Most controllers allow it and it is rarely wise. Forcing continues past errors and can write incorrect data across the replacement, which compounds rather than resolves.
It completed using bad data from a failing member. The original disks are still the better source, so do not wipe them — send the full set including the one you replaced.
No. Changing anything mid-incident removes information about the array's state. Take it out of service as it stands.
Urgent in the sense that leaving it powered risks more, and not urgent in the sense that a powered-down array is stable indefinitely. Switch it off and then take your time.
From £500 +VAT after the free diagnostic, fixed in writing.