My Plex server on Unraid went down hard this week. No warning, no gradual degradation — just stopped working. After a few restarts produced the same result, I dug into the logs and found this:

Starting Plex Media Server.
Error: Unable to set up server: std::exception (N2DB9ExceptionE)
Stopping Plex Media Server

Classic corrupted database. Here’s how I fixed it, what I tried first, and why a simple backup restore ended up being the answer.

First Attempt: Manual Database Repair

The official Plex documentation isn’t great for Docker setups on Unraid. The repair process requires SQLite — which is bundled with Plex — but you need Plex stopped while running the commands. That means using the container’s console rather than running things from the Unraid terminal.

Stop the Plex Service Inside the Container

In Unraid, go to Docker, click the Plex container icon, and open the console.

For the older plexinc/pms-docker container:

./plex_service.sh -d

For the newer lscr.io/linuxserver/plex container:

cd /var/run/s6-rc/servicedirs/
s6-svc -d svc-plex

Navigate to the Database Directory

cd "/config/Library/Application Support/Plex Media Server/Plug-in Support/Databases/"

Back Up Before Touching Anything

cp com.plexapp.plugins.library.db com.plexapp.plugins.library.db.bak

Don’t skip this. Even if the database is corrupted, keep the original.

Run an Integrity Check

"/usr/lib/plexmediaserver/Plex Media Server" --sqlite com.plexapp.plugins.library.db "PRAGMA integrity_check"

If this returns anything other than ok, the database has errors.

Attempt a Quick Recovery

"/usr/lib/plexmediaserver/Plex Media Server" --sqlite com.plexapp.plugins.library.db ".output recover.out" ".recover"

This tries to salvage data from a corrupted database. In my case, it wasn’t enough.

Second Attempt: Full Manual Dump and Rebuild

When the quick recovery doesn’t cut it, the next step is dumping the database contents and rebuilding from the dump.

Dump the Database

"/usr/lib/plexmediaserver/Plex Media Server" --sqlite com.plexapp.plugins.library.db .dump > dump.sql

Remove the Corrupted File

rm com.plexapp.plugins.library.db

Rebuild From the Dump

"/usr/lib/plexmediaserver/Plex Media Server" --sqlite com.plexapp.plugins.library.db < dump.sql

Fix File Ownership

For plexinc/pms-docker:

chown plex:plex com.plexapp.plugins.library.db

For lscr.io/linuxserver/plex:

chown abc:users com.plexapp.plugins.library.db

This also didn’t work. The rebuilt database came back as 0 bytes — the original was too far gone to recover anything useful from.

The Actual Fix: Restoring a Plex Database Backup

Here’s where the backup I’d set up months ago saved the day. Plex has a built-in scheduled backup — it’s buried in Settings > Scheduled Tasks and it’s off by default. I’d enabled it and set it to run every 3 days.

My database directory listed several clean backups:

-rw-r--r-- 1 plex users  81M Mar 18 03:16 com.plexapp.plugins.library.db-2026-03-18
-rw-r--r-- 1 plex users  81M Mar 21 02:06 com.plexapp.plugins.library.db-2026-03-21
-rw-r--r-- 1 plex users  81M Mar 24 02:06 com.plexapp.plugins.library.db-2026-03-24

I grabbed the most recent one and ran it through an integrity check first.

Restore the Backup

cp com.plexapp.plugins.library.db-2026-03-24 com.plexapp.plugins.library.db
chown plex:users com.plexapp.plugins.library.db

Verify It’s Clean

"/usr/lib/plexmediaserver/Plex Media Server" --sqlite com.plexapp.plugins.library.db "PRAGMA integrity_check"

Output:

ok

That’s it. Restarted Plex, it came straight up, library intact.

Enable Plex Database Backups Now

If you’re not already running scheduled backups, turn them on before you need them.

In Plex Web, go to Settings → Scheduled Tasks and enable Backup database every X days. Three days is a reasonable interval — enough history to roll back past corruption without eating too much storage. Each backup is a full copy of the database, so on an 80MB database that’s around 240MB of backup files at any given time. Cheap insurance.

The backups land in the same directory as the live database and are named with a datestamp, making it obvious which is which.

When Backups Aren’t an Option

If you’re in a situation where backups aren’t available and the manual repair steps above haven’t worked, there’s a community tool worth knowing about: DBRepair by ChuckPa. It’s a more thorough repair utility built specifically for Plex databases, and there’s an Unraid-specific Docker support thread on the Unraid forums. Useful for large, complex libraries shared across multiple users where a backup restore isn’t an option.

For most home setups though, a clean backup beats every repair tool available.

Plex Database Repair on Unraid: Summary

The manual SQLite repair process works in some cases, but if the database is badly corrupted, the dump will produce a 0-byte file and you’re stuck. The real lesson here is to have the backup in place before it’s needed.

Steps that fixed it for me:

1. Confirmed the database was corrupted via the error log
2. Attempted quick recovery and full dump-rebuild — neither worked
3. Restored from a 1-day-old Plex scheduled backup
4. Ran PRAGMA integrity_check to confirm the restored file was clean
5. Restarted Plex — library came up fully intact

Enable database backups in Plex’s Scheduled Tasks. It’s free, it’s built-in, and it’s the only thing that will save you when a repair fails.

For more Docker and self-hosting guides on Unraid, check out the Unraid category on BlackmoreOps.