Server Not Detecting a Drive: What to Check and In What Order
A drive that has disappeared is usually still there. This covers the order to check things in, from the causes that cost nothing to fix to the ones that mean a replacement.
Start with what the controller says
Before touching anything physical, look at what the system reports.
The RAID controller or management interface lists every bay and what it sees there. A drive showing as failed, foreign, or unconfigured is a different problem from a bay showing as empty.
Three states and what each means.
Bay shows empty. The controller sees nothing. Physical connection, backplane or a genuinely dead drive.
Drive present but marked failed. The controller sees it and has rejected it. Could be a real failure, could be the drive dropped out during an error recovery pause.
Drive present but foreign or unconfigured. The controller sees a drive carrying a different array configuration. Common when moving drives between machines, and importable rather than broken.
Our controller guide covers reading these.
The free checks, in order
Reseat the drive. Pull it and push it home firmly. Carrier connectors work loose from vibration and from rack work. This fixes a meaningful share of missing drives and costs nothing.
Try a different bay, if one is free. If the drive appears in another bay, the problem is the backplane or that bay’s connector, not the drive.
Check the backplane cable where the chassis is accessible. A partially seated SAS cable can take a whole group of bays offline at once — if four bays disappeared together, this is almost certainly why. Our cables guide covers the connector types.
Check the drive spins up. Hold it as you insert it; you can feel a drive spin. No spin usually means a dead drive or no power to the bay.
If the drive is marked failed
Read the health data before replacing it.
A drive with growing reallocated sectors or uncorrectable errors has genuinely failed. Replace it. Our drive failure guide covers the values.
A drive with clean health values but rising command timeouts frequently has not failed at all. It stopped responding long enough for the controller to drop it, which happens when a desktop-class drive does extended error recovery in an array. The drive is fine; it is the wrong class of drive. Our drive class guide covers why this happens.
Where several drives across different bays show timeouts, suspect the cable, backplane or expander rather than the drives.
If a whole group of bays disappeared
Drives do not fail in groups. Infrastructure does.
Four causes. A backplane cable serving those bays. A SAS expander failing. A power connector to that section of the backplane. Or the controller itself losing a port or channel.
The pattern tells you which: contiguous bays point to a cable or backplane section; every bay at once points to the controller.
When the drive is fine but shows as foreign
The drive carries array metadata from another controller or another array.
Controllers offer an import option for this, which reads the existing configuration, and a clear option, which discards it. Import preserves the data; clear destroys it. Be certain which you want before choosing, and take a backup first if the data matters.
This is also worth knowing when buying used drives: a drive from a decommissioned array frequently arrives with foreign configuration on it, and that is normal rather than a fault.
After it is fixed
Two things.
Let the rebuild complete before doing anything else to the array. A rebuild loads every remaining drive heavily, and this is the moment a second failure happens. Our RAID guide covers why.
Record what happened. A drive that dropped once and came back is worth watching; a bay that has dropped drives twice is a backplane problem waiting to be diagnosed properly. Our monitoring guide covers spotting patterns.
Frequently asked questions
My server is not detecting a drive. What do I check first?
What the controller reports for that bay, then reseat the drive. Carrier connectors work loose from vibration and rack work, and reseating fixes a meaningful share of missing drives at no cost.
Several drives disappeared at once. Are they all dead?
Almost certainly not. Drives do not fail in groups; infrastructure does. Suspect a backplane cable, a SAS expander, a power connector to that backplane section, or the controller. Contiguous bays point to a cable; every bay points to the controller.
What does foreign configuration mean on a drive?
The drive carries array metadata from another controller or array. Controllers offer import, which preserves the data, and clear, which destroys it. Be certain which you want. It is normal on drives from decommissioned arrays.
The controller marked a drive failed but it seems fine. Why?
Check whether health values are clean and command timeouts are rising. A drive that stopped responding long enough to be dropped, with no sector errors, is frequently a desktop-class drive doing extended error recovery in an array. The drive is fine; the class is wrong.
How do I tell if it is the drive or the bay?
Try the drive in a different bay. If it appears there, the problem is the backplane or that bay’s connector rather than the drive.
Can I do anything to the array while it is rebuilding?
Let the rebuild complete first. It loads every remaining drive heavily, and this is exactly when a second failure happens.
Tell us the server model and what the controller reports, and we will help you work out whether you need a drive, a cable or a backplane.
