4.1.2. Troubleshooting Storage (Drives and RAID)
💡 First Principle: To diagnose storage failures, you must interpret OS errors (e.g., "Bootable device not found"), listen for physical cues (clicking), and check status indicators (S.M.A.R.T., RAID controller) to determine if the issue is logical or physical.
Scenario: A user reports their computer is extremely slow to boot and open files. You hear a faint, repetitive clicking sound coming from the case. You immediately form the theory that the mechanical hard drive is physically failing.
Visual: Internal view of a hard disk drive with the cover removed — the spinning platters store data magnetically, and the read/write head on the actuator arm moves across them. A clicking sound from this mechanism indicates imminent physical failure.
Technician's Diagnostic Path:
- Primary Action: Inform the user to stop using the computer to prevent further damage. Your immediate priority is data preservation, not repair.
- Plan for Data Recovery:
- Action: Connect the failing drive to a different computer as a secondary drive (e.g., via a USB-to-SATA adapter).
- Action: Attempt to copy the user's critical data to a new, healthy drive.
- Plan for Repair:
- Action: Replace the failing HDD with a new SSD for a significant performance improvement.
- Action: Perform a clean installation of the operating system on the new SSD and restore the user's backed-up data.
⚠️ Common Pitfall: Attempting to "fix" a physically failing hard drive with software tools. If a drive is making physical noises, the priority is to back up the data before it fails completely, not to run disk checkers.
Key Trade-Offs:
- Data Recovery Software vs. Professional Service: For critical data on a failed drive, professional data recovery services have specialized tools and clean rooms but are very expensive. Software tools are a last resort and can sometimes make the problem worse.
Storage Symptom Quick Reference:
| Symptom | Most likely cause | First action |
|---|---|---|
| Clicking or grinding from an HDD, rising S.M.A.R.T. reallocated sectors, repeated file corruption | Failing mechanical drive | Back up first (power off and plan data recovery if it will not load); do not run disk-repair tools on a drive making physical noise |
| Constant 100% disk usage on a mechanical HDD, even at idle | A failing drive retrying bad sectors; if S.M.A.R.T. is healthy, the HDD itself is the bottleneck (too little RAM drives the disk through paging mainly when many apps are open) | Check S.M.A.R.T.; back up and replace a failing drive; replace a healthy but slow HDD with an SSD (defragmenting or reinstalling Windows gives no significant gain) |
| Drive suddenly goes read-only, with no clicking | An SSD that has exhausted its write endurance | Copy the data off and replace the SSD |
| "Bootable device not found" and the drive is not listed in UEFI/BIOS | Loose or faulty SATA data or power cable, or a failed drive | Power off and reseat the cables, try another cable/port, test the drive in another system |
| "OS not found" but the drive is listed in UEFI/BIOS | Wrong boot order, or a corrupted boot sector/MBR/BCD | Check the boot order, then repair the boot files |
| New SSD installed and properly connected, "Bootable device not found" | Boot order not pointing at the new drive | Check the BIOS/UEFI boot order first |
| New NVMe SSD properly seated but missing from UEFI/BIOS | Storage controller mode (AHCI/RAID/NVMe) set wrong | Check the storage controller mode setting |
| KERNEL_DATA_INPAGE_ERROR blue screens | Storage drive (or its cable/controller) failing | Check S.M.A.R.T. and the cabling, then replace the drive |
Recommended order for "Boot device not found": (1) check whether the drive is detected in BIOS/UEFI; (2) verify the boot order; (3) power off and reseat the drive's data and power cables; (4) test the drive in another system or with diagnostics; (5) repair the boot files or replace the failed drive.
RAID status after one drive fails ("degraded" means the array still works but has lost its redundancy):
| RAID level | Status | What to do |
|---|---|---|
| RAID 0 | The whole array is lost (no redundancy) | Replace the drive and restore from backup |
| RAID 1 | Degraded but operational; the surviving drive of the mirror holds all the data | Replace the failed drive and rebuild |
| RAID 5 | Degraded but operational (missing data is rebuilt from parity, so it runs slower); a second failure would lose the array | Replace the failed drive and rebuild promptly; confirm backups |
| RAID 10 | Degraded but operational, and capacity is unchanged; it survives as long as both drives of one mirrored pair do not fail | Replace the failed drive and rebuild |
Storage tools: S.M.A.R.T. is the drive's built-in self-monitoring; check it to assess health proactively, before a failure. CHKDSK C: /f checks and repairs file-system integrity (it cannot fix a mechanically failing drive). Defragmentation only helps HDD performance. diskpart manages partitions and format erases a volume.
Reflection Question: In a RAID 5 array, one drive fails. What is the status of the array, what is the immediate risk, and what is the first step you should take?
