RAID 5 Degraded Recovery — Step-by-Step Guide

How RAID 5 Works — The Degraded State Explained

RAID 5 distributes data and parity information across three or more drives. For every stripe of data written across N drives, one drive's worth of capacity is used for parity — a mathematical value that allows any single drive's data to be reconstructed from the remaining drives.

When one drive fails, the array enters degraded mode. This is a critical but not yet fatal state. Here's what it means exactly:

• All data is still accessible — the controller reconstructs the failed drive's data on-the-fly using parity
• Read performance drops because every read now requires parity calculation from all remaining drives
• Write performance drops significantly — every write must also update parity across all drives
• Zero fault tolerance remains — one more drive failure means total array loss

The degraded state is a race condition: you need to either replace the failed drive and complete a rebuild, or back up all data — before another drive fails. RAID 5 drives in a NAS that has been running 24/7 for 3-5 years are all near their failure threshold simultaneously. When one fails, the others are statistically close behind.

First 5 Steps When Your RAID 5 Goes Degraded

Follow this sequence exactly when your NAS shows a degraded RAID 5 notification.

Step 1: Do not power off the NAS unexpectedly

An unexpected power interruption while the RAID is degraded can corrupt the array's RAID metadata, turning a degraded array into an offline array. If you need to power off (for transport or drive removal), use the proper NAS shutdown sequence through the management interface. Never just unplug the power cable.

Step 2: Identify which drive failed

In your NAS management interface, check which specific drive bay is showing as failed. Note the drive model, serial number, and which bay number. In Synology DSM: Storage Manager → Storage Pool → Physical Disks. In QNAP: Storage & Snapshots → Storage → Disks.

Step 3: Check S.M.A.R.T. on all remaining drives immediately

This is the most critical step most NAS admins skip. Before doing anything else, run S.M.A.R.T. tests on all remaining drives in the array. In Synology DSM: Storage Manager → HDD/SSD → select each drive → Health Info → Extended S.M.A.R.T. Test. If any remaining drive shows "Caution" or "Bad" status, high reallocated sector counts, or pending sectors — do NOT rebuild. Back up first.

Step 4: Check available backup destinations

Can you copy the NAS data to another location before rebuilding? An external drive, Backblaze B2, S3, or another NAS? If yes, start a backup immediately — before touching any drive replacement or rebuild. Even a partial backup of the most critical folders is better than no backup.

Step 5: Assess rebuild vs backup path

If all remaining drives are S.M.A.R.T. healthy and you have backup capacity: replace the failed drive, let the array rebuild, then implement a backup strategy going forward. If any remaining drive shows S.M.A.R.T. warnings, or if you don't have backup capacity: call IT Cares at 1 (888) 711-9428 before rebuilding.

Why You Must NOT Rebuild Immediately Without Imaging

The NAS management interface makes it look simple: replace the failed drive, click "Repair" or "Rebuild," and wait. This is fine if your surviving drives are perfectly healthy. It becomes catastrophic if they are not.

RAID 5 rebuild reads every single sector of every surviving drive. For a 3-drive RAID with 4 TB drives, that's 8 TB of sequential reads under sustained load. Consumer and even NAS-grade drives (WD Red, Seagate IronWolf) have a published Unrecoverable Read Error (URE) rate of 1 error per 10^14 bits, which equals roughly one URE per 12.5 TB of data read.

For a 3-drive RAID 5 with 4TB drives each, the rebuild reads 8 TB — giving a statistical probability of URE during rebuild that grows with array size. Larger arrays (5-drive, 8-drive) have near-certain URE probability during rebuild. A single URE encountered during rebuild causes the controller to declare the rebuild failed and take the array offline.

Synology DSM Degraded RAID Recovery

Synology uses its own RAID implementation called SHR (Synology Hybrid RAID) in addition to standard RAID 5/6. Here is the specific approach for each.

Synology SHR/RAID 5 — drive shows as failed

When DSM marks a drive as failed: go to Storage Manager → Storage Pool → look for the red drive indicator. Before replacing anything, run extended S.M.A.R.T. tests on all other drives. If all pass: remove the failed drive, insert a new drive of equal or larger capacity, and let DSM run the repair. DSM will rebuild automatically and show a progress percentage. Do NOT shut down during rebuild.

Synology RAID — array shows as "Crashed"

If DSM shows "Storage Pool Crashed" rather than "Degraded," two or more drives have failed, or the rebuild attempt failed. Do NOT attempt another repair from the DSM interface — this can make recovery harder. Power off cleanly (shutdown through DSM UI if possible), remove all drives, label them (Bay 1, Bay 2, Bay 3...) and call IT Cares. Drive order and bay assignment are critical for professional RAID recovery.

Synology boot failure after array issue

If the Synology NAS won't boot after a RAID issue: connect your drives to a Linux system with mdadm (Synology SHR uses mdadm internally) and attempt to assemble the array in read-only mode. This allows data extraction without triggering a rebuild. This is an advanced procedure — IT Cares can guide you through it or handle it remotely.

QNAP Degraded RAID Recovery

QNAP uses their QTS (or QuTS Hero for ZFS-based systems) management interface. RAID 5 in QNAP behaves similarly to Synology but with different management screens.

QNAP RAID 5 degraded — single drive failure

In Storage & Snapshots → Storage → Storage/Snapshots, find your volume. A degraded state shows with a yellow warning. Check each disk's S.M.A.R.T. in Disk Management → Disks/VJBOD before replacing. If all remaining drives are healthy: hot-swap the failed drive (QNAP supports hot-swap on most models) and QNAP will automatically start the rebuild.

QNAP "Data volume degraded/failed" — multiple drive issues

If QNAP shows the volume as failed (red) rather than degraded (yellow): the array is offline. Do NOT use QNAP's "Recover" function without understanding what it will do — in some QNAP firmware versions, "Recover" attempts a forced rebuild that overwrites parity information. Power off cleanly, remove and label all drives, and ship to IT Cares for professional recovery.

Windows Storage Spaces RAID Recovery

Windows Storage Spaces provides software RAID functionality for Windows 10/11 and Windows Server. It supports parity (RAID 5 equivalent) and mirroring (RAID 1) configurations.

Storage Spaces degraded — drive failed

Open Settings → System → Storage → Storage Spaces (Windows 10/11) or Server Manager → File and Storage Services (Server). A failed drive shows in the pool with a warning. Before removing: check the Event Viewer (Windows Logs → System) for disk errors on remaining drives. Replace the failed drive and Windows will automatically repair the space.

Storage Spaces — pool shows as "degraded" with I/O errors

If the pool is degraded with I/O errors logged for multiple drives: do not attempt repair from the UI. Copy critical data immediately if any data is accessible. Then call IT Cares. Windows Storage Spaces metadata is stored on the drives themselves — professional recovery tools can read and reconstruct this metadata without the Windows installation that created the space.

Linux mdadm Software RAID Recovery

mdadm (Multiple Device Administration) is the standard Linux software RAID tool. Synology NAS devices use mdadm internally, and many Canadian server deployments use mdadm RAID 5 on Ubuntu or CentOS servers.

Checking array status

Check current state: cat /proc/mdstat — shows all md devices, their state (active/degraded), and recovery progress if running. mdadm --detail /dev/md0 — shows detailed array information including which drives are active, which are failed, and the RAID configuration.

mdadm RAID 5 — one drive failed, array degraded

If /proc/mdstat shows the array as "active, degraded" with [U_U] or similar pattern (U = Up, _ = Down):

1. First, image all surviving drives before anything else: ddrescue /dev/sdb /path/to/backup/drive_sdb.img /path/to/backup/drive_sdb.log
2. Run S.M.A.R.T. on all remaining drives: smartctl -a /dev/sdb
3. Only if all surviving drives are S.M.A.R.T. healthy: add new drive mdadm /dev/md0 --add /dev/sdd
4. Monitor rebuild: watch cat /proc/mdstat

mdadm — array will not assemble

If mdadm --assemble fails to bring up the array: try force-assemble with fewer drives: mdadm --assemble --force /dev/md0 /dev/sdb /dev/sdc (use working drives only). If this succeeds in read-only mode, mount the array as read-only (mount -o ro /dev/md0 /mnt/recovery) and copy data immediately. Do NOT write to the array in this state.

When Rebuild Goes Wrong — What IT Cares Does

We receive calls weekly from clients who started a RAID rebuild, had it fail partway through, and now have a worse situation than when they started. Here is how professional RAID recovery differs from what you can do with the NAS management interface.

Virtual RAID reconstruction without a controller

IT Cares removes each drive from the NAS and connects them individually to workstations running professional RAID recovery software (R-Studio, UFS Explorer RAID Recovery, ReclaiMe). These tools can virtually reconstruct the RAID without requiring the original controller or NAS system.

The software analyzes each drive to determine: the stripe size (typically 64 KB or 128 KB), the drive order within the array, the parity rotation method (left-async, right-sync, etc.), and the RAID metadata (LVM, mdadm superblock, or Synology-specific). Once the virtual RAID is reconstructed in software, it appears as a single logical volume from which files can be extracted.

Handling drives with bad sectors or UREs

When one or more surviving drives have bad sectors, professional tools use fault-tolerant imaging: ddrescue images the drive in multiple passes, skipping and logging bad sectors on the first pass, then returning to retry them. The RAID reconstruction tools then handle the gaps intelligently — using parity from other drives to fill in missing data where possible.

Recovering from a partially rebuilt array

When the NAS started a rebuild and a second drive failed partway through: the replacement drive now contains a partial rebuild (some stripes rebuilt, some not). Professional tools identify which stripes were completed versus which are still in the original partial-failure state, and reconstruct accordingly. This is complex but successful in most cases.

Understanding UREs — Why Large RAID Arrays Are at Risk

A URE (Unrecoverable Read Error) is a permanent sector failure that cannot be read after the drive's maximum retry count. Consumer drives are rated at 10^14 URE rate — one error per approximately 12.5 TB of sequential reads.

Enterprise drives (WD Gold, Seagate Exos) are rated at 10^15 URE rate — 10x better. For NAS arrays larger than 4 drives or with drives larger than 4 TB, using NAS-grade or enterprise drives with the lower URE rate significantly reduces rebuild risk.

Prevention — How to Avoid Future RAID Failures

The best RAID 5 recovery is the one you don't need. Here are the practices that keep RAID arrays reliable across Canada's small business and home office environments.

Replace drives proactively every 3-4 years

Hard drives in 24/7 NAS environments accumulate 26,000–35,000 power-on hours in 3-4 years. Most manufacturers rate desktop drives for 8,760 hours/year (1-shift use) — NAS use runs at 2-3x the rated workload. Replace the full drive set every 3-4 years regardless of whether drives show failure signs. Staggering replacement (replace one drive every 6-12 months) keeps the array running during replacement.

Monitor S.M.A.R.T. data weekly

Enable email notifications in your NAS management interface for S.M.A.R.T. warnings. In Synology DSM: Control Panel → Notification → Email. In QNAP: System → Notification Center. Set alerts for Reallocated Sectors Count increasing, Pending Sectors, and Uncorrectable Errors. These metrics typically warn you 30-90 days before outright drive failure.

RAID is not a backup

RAID 5 protects against single drive failure only. It does not protect against: accidental file deletion, ransomware encryption (which will encrypt your RAID volume), NAS firmware corruption, flooding or fire, or theft. Implement the 3-2-1 rule: 3 copies, 2 media types, 1 off-site (Backblaze B2 is $6/TB/month and works natively with Synology's Cloud Sync and QNAP's Hybrid Backup Sync).

Keep drives from the same batch separate

Drives purchased together (same batch, same manufacturing date) fail together — they have the same number of hours, the same thermal history. When building a new RAID, buy drives from different batches or different vendors to reduce the statistical probability of simultaneous failure.

Frequently Asked Questions

Services: RAID Data Recovery · Data Recovery Hub · External Drive Recovery · Toronto · Brampton