A failed disk in a redundant RAID volume is urgent but usually recoverable if you replace the right bay, use a compatible drive, and let the array rebuild cleanly. Pull the wrong disk and you turn a degraded array into data loss.

This is the replacement path. For why RAID is not a backup, see /what-is-raid-drive-failure-protection-explained. Level tradeoffs live at /raid-levels-explained-which-one-for-your-nas. Prefer CMR NAS drives (/wiki/cmr); avoid SMR rebuild traps (/wiki/smr). Pair the fix with a real 3-2-1 plan at /3-2-1-backup-strategy-nas.

A typical 2-bay host such as the Synology DiskStation DS225+ shows bay status in DSM Storage Manager and on the front LEDs, which is what you use to avoid guessing trays.

What you need

  • NAS admin access (DSM, QTS/QuTS, TrueNAS, or similar) and a working Storage Manager view of the volume.
  • A replacement HDD: same SATA/SAS interface, same or larger capacity than the failed member, preferably CMR NAS-class (see /wiki/cmr).
  • Antistatic handling, the correct drive tray/screws for your chassis, and a calm window of several hours for rebuild I/O.
  • A known-good backup if the array is already past its redundancy limit.

A like-for-like CMR pick such as the WD Red Plus 4TB matches common home NAS bays. If that SKU is unavailable, a NAS alternate like the Seagate IronWolf 4TB is the usual substitute.

For what a NAS is in the first place, see /what-is-a-nas-network-storage-explained.

1. Confirm which bay failed

  1. Open Storage Manager (Synology DSM) or Storage & Snapshots / Storage Manager (QNAP). Note the volume state: Degraded, Warning, or Critical.
  2. Identify the disk marked Failed, Crashed, or Abnormal. Record the bay number, model, capacity, and serial if shown.
  3. Match the UI bay to the physical chassis LED. Many enclosures light amber/red on the failed tray; do not trust memory of which bay you installed last year.
  4. Optional: pull SMART details for that disk only. Reallocated sectors, pending sectors, or offline uncorrectable counts often explain the failure and are useful early-warning signals on the remaining members.

2. Order a compatible replacement

  1. Match interface (almost always SATA 6Gb/s for home NAS HDDs).
  2. Match form factor (3.5 inch in desktop NAS trays; 2.5 inch only if that bay was 2.5 inch).
  3. Capacity must be equal or larger than the failed member. Never install a smaller disk into an existing RAID set.
  4. Prefer CMR NAS-class drives (WD Red Plus, IronWolf, and similar). SMR desktop drives can make rebuilds painfully slow or fail under array stress. Background on CMR vs SMR: /wiki/cmr and /wiki/smr.
  5. Same brand/model is convenient for firmware quirks, but not required. Same capacity class and CMR NAS rating matter more.

3. Power and hot-swap notes by chassis

  • Hot-swap trays (most Synology/QNAP desktop units with tool-less sleds): you can usually deactivate the failed disk in Storage Manager, unlock the tray, and pull it while the NAS stays powered. Follow the vendor dialog; do not yank a tray mid-write without deactivating when the UI offers that step.
  • Non-hot-swap or screw-in cages: shut down cleanly from the OS, power off, swap the drive, power on, then start repair.
  • USB/eSATA DIY enclosures: treat as non-hot-swap unless the vendor explicitly documents hot-plug for that bay.
  • Keep the NAS on a UPS if you have one. Dirty power during rebuild is a common way to lose a second disk mid-recovery. UPS shutdown patterns are covered separately on the site when you need that path.

4. Deactivate and remove only the failed bay

  1. In Storage Manager, select the failed disk and choose Deactivate, Disable, or the vendor equivalent if offered.
  2. Wait until the UI shows the bay as deactivated or empty-ready.
  3. Unlock the matching physical tray. Double-check bay number against the LED and the UI one more time.
  4. Slide the failed drive out. Set it aside labeled "failed" with the date. Do not reuse it in another array without a full wipe and SMART check, and do not trust it for sole copies of data.

5. Insert the new drive and start rebuild

  1. Mount the new HDD in the empty tray (orientation and SATA connector facing the backplane as the failed disk did).
  2. Slide the tray in until it seats. Lock the latch.
  3. In Storage Manager, the new disk should appear as uninitialized or available. Select the degraded storage pool/volume and choose Repair, Rebuild, or Replace disk.
  4. Confirm the target is the new bay only. Start the job.
  5. Leave the NAS online. Rebuilds are heavy sequential and random I/O; expect slower shares until the job finishes.

6. Rebuild time, load, and what "healthy" means again

Rebuild duration scales with capacity, RAID level, how full the volume is, and concurrent client load. A few-terabyte home RAID 1 can finish in hours; multi-disk RAID 5/6 on larger disks can take overnight or longer. DSM and QNAP show percent complete and estimated time; those estimates move.

  • Avoid large downloads, Plex library scans, and Hyper Backup jobs on the same volume during the first half of a rebuild if the NAS is already struggling.
  • Do not reboot casually mid-rebuild unless the vendor UI says it is safe to resume.
  • When status returns to Healthy / Normal, spot-check a few critical folders and confirm remaining disks show clean SMART.

SMART early warnings (before the next failure)

RAID rebuilds stress every surviving member. After recovery:

  • Enable email or push alerts for disk warnings in Control Panel / Notification.
  • Watch reallocated and pending sector counts on the remaining drives.
  • Schedule a data scrub / consistency check on a quiet night once the array is healthy again.
  • Keep a spare CMR drive of the array capacity on the shelf if uptime matters.

When to stop and restore from backup instead

Stop DIY repair and restore from backup when:

  • More disks have failed than the RAID level can tolerate (second failure in RAID 5, third in RAID 6, both members in RAID 1).
  • The volume mounts read-only, missing, or with filesystem errors after the swap.
  • Rebuild fails repeatedly on the new disk (bad replacement, wrong capacity, or deeper pool corruption).

RAID is redundancy for a single-disk fault window, not a substitute for offsite copies. Layout guidance: /3-2-1-backup-strategy-nas. Background on failure modes: /what-is-raid-drive-failure-protection-explained. Level choice after you recover: /raid-levels-explained-which-one-for-your-nas.

Bottom line

Identify the failed bay in software and on the LED, buy a same-or-larger CMR NAS drive, deactivate only that tray, insert the replacement, and start Repair/Rebuild. Never pull a healthy member from a degraded array. Prefer CMR over SMR for rebuild-friendly members, and keep a real backup so a second failure is inconvenient rather than catastrophic.