clear_drive_error

Clear Synology DSM drive error history so Storage Manager will accept a disk again.

Author: Zilloww
Inspired by 007revad/Synology_clear_drive_error.

---

When this helps

DSM can mark a disk Critical and refuse Repair, even when S.M.A.R.T. looks fine.

Typical case:

• Storage pool is degraded
• One disk is Critical / kicked from the array
• Raw S.M.A.R.T. (5 / 197 / 198 / 199) is clean
• Repair says the disk does not meet requirements (must be Healthy)

DSM stores that judgment in sqlite DBs under /var/log/synolog/. This script deletes all history for the selected serial across disk / test / health DBs. It does not fix hardware.

---

Before you run it

Worth trying only if the disk looks like a false positive / transient glitch, not a dying drive.

Prefer yes if:

• Reallocated / Pending / Uncorrectable / CRC raw values are 0 (or stable)
• Extended S.M.A.R.T. passes (or you plan to run one after clearing)
• You understand you are wiping DSM’s memory of past errors for that serial

Prefer no if:

• SMART is already bad and getting worse
• The disk keeps dropping / reconnecting
• You have no backup of data you care about
• You are not comfortable with root SSH on the NAS

Risk reminder (RAID 5 / SHR-1): while degraded or rebuilding you have no redundancy. A second disk failure can kill the volume. Rebuilding on the same disk vs a new one: same catastrophe risk; same disk has a higher chance of rebuild abort / re-kick.

---

Install

scp -O clear_drive_error.sh user@NAS_IP:~/scripts/

ssh user@NAS_IP
sudo -i
chmod +x /var/services/homes/USER/scripts/clear_drive_error.sh

Needs: SSH, admin/root, sqlite3.

---

Usage

./clear_drive_error.sh --list
./clear_drive_error.sh --dry-run
./clear_drive_error.sh

Pick the drive from the menu (model|serial). Use --yes to skip the final confirmation.

Flag	Meaning
-l, --list	List drives with err/warning history
-n, --dry-run	Preview only
-y, --yes	No confirmation prompt

Always clears everything for that serial. Backups are written as *.bak.<timestamp> before any write.

---

Full repair flow

1. Confirm pool degraded, note slot + serial, check SMART raw values.
2. Clear history

   ./clear_drive_error.sh --dry-run   # pick the drive, preview
   ./clear_drive_error.sh            # pick the drive, confirm

3. Reseat cleanly
   • Deactivate if the Action exists (often missing when already Critical — fine)
   • Pull the disk
   • Reboot without that disk
   • Reinsert in the same bay
4. Expect Healthy / Not initialized / unused in Storage Manager.
5. Soak (recommended) before Repair:
   • Leave the disk in, do not Repair yet
   • Run an extended S.M.A.R.T. test
   • Next day, check for new errors:

     sqlite3 /var/log/synolog/.SYNODISKDB \
       "SELECT datetime(time,'unixepoch','localtime'),level,msg FROM logs \
        WHERE serial='SERIAL' AND level IN ('err','warning') \
        ORDER BY time DESC LIMIT 30;"

6. Repair the storage pool only if still Healthy and no new unc / ioerr / timeout / idnf.
7. After rebuild: wait for Healthy, then run a data scrub.

If Repair still refuses: repeat pull → reboot empty → reinsert. Last resort (experts only): manual mdadm --add — easy to brick the array if you pick the wrong device.

---

Disclaimer

Use at your own risk. This edits internal DSM databases, is not an official Synology tool, and will not save a failing drive. Wrong target, bad timing, or a rebuild you do not understand can mean downtime or data loss.

You should know what a degraded RAID means and ideally have a backup of anything you cannot lose before you click Repair.

If unsure: stop, replace the disk, or get help.