Upgrade to 102 Overwrote Database

[Revali25]

Description

I updated my docker instance of trilium-next to the most recent version and in doing so the data apparently has been deleted, it thinks that the database has not been intialized. I can't find the data directory on my disk anymore, and the only document.db that came back in a find search was from a snapshot, which I'm at least happy I have that. I can't find any of the backups anywhere else on the disk, it's like they vanished. I don't have logs either because I can't get into the app without setting it up as a new instance. I was also able to pull a document.db from TriliumDroid too, and I'm not sure where I even should put that now that I don't see any data directories.

TriliumNext Version

102

What operating system are you using?

Other Linux

What is your setup?

Local + server sync

Operating System Version

Raspbian 13

Error logs

No response

[dosubot[bot]]

Sorry to hear about your data loss! This sounds like it could be related to a known issue where Docker mount paths changed between versions (historically from /root/trilium-data to /home/node/trilium-data), causing data to appear "missing" after upgrades.

To restore from your snapshot/TriliumDroid backup:

1. Stop the Docker container completely
2. Place your document.db in ~/trilium-data/ on your host machine (this maps to /home/node/trilium-data/ inside the container)
3. Important: Delete any existing document.db-wal and document.db-shm files in that directory - they must be removed to avoid restart loops
4. Set permissions: chmod 600 ~/trilium-data/document.db
5. Verify your Docker volume mount looks like:

   volumes:
     - ~/trilium-data:/home/node/trilium-data

6. Start the container

Finding your old data: Check if your original data might still exist at /root/trilium-data/ or another mount path from an older configuration. Your backups would normally be in ~/trilium-data/backup/ with names like backup-daily.db, backup-weekly.db, and backup-monthly.db.

For TriliumDroid backup: The same restoration steps apply, but note that sync version compatibility matters - if your TriliumDroid was on a different version, you may encounter issues.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[Revali25]

I usually follow the default directory structures for things like this because then troubleshooting becomes easier, and my home directory doesn't have a trilium-data folder anymore it looks like? I can't even remember what the directories were structured like, because I usually don't think about it, it just works. If I just put the document.db file in a new folder ~/trilium-data and map the volume mount to ~/trilium-data then it should still work right?

[dosubot[bot]]

Yes, that should work! Create ~/trilium-data, put your document.db in there, and map the volume to /home/node/trilium-data in the container:

volumes:
  - ~/trilium-data:/home/node/trilium-data

This is the default configuration in docker-compose.yml. Just make sure to delete any document.db-wal and document.db-shm files if they exist alongside your backup, and Trilium will handle any necessary migrations on startup.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[eliandoran]

Were you able to solve your issue?
It sounds like there was something wrong with your volume mount in Docker. When the image tag got changed, the local changes were lost.

[eliandoran]

@Revali25 , yes, the easiest to restore the backup for now would be to do it locally.
If you get "Sync in progress", that means your database is most likely incomplete (a snapshot from an uninitialized instance). Is this the backup from TriliumDroid?

Could you show me the content of your docker.compose.yml file?

[Revali25]

@eliandoran Yes the backup from TriliumDroid is the one that's showing Sync in progress

The docker compose file is this:

services:
trilium:
container_name: trilium
image: triliumnext/trilium:latest
environment:
- TRILIUM_PUBLIC_URL=https://trilium.[DOMAIN].net
volumes:
- ~/trilium-data:/root/trilium-data
restart: unless-stopped
networks:
- proxy_network

volumes:
trilium_data:
external: true
name: cloud_stack_trilium_data

networks:
proxy_network:
external: true

[Revali25]

I did realize that the volume was mapped incorrectly to /root/trilium-data; I fixed that and now the same sync error is happening. I'm assuming that you're right, the database is probably from an uninitialized snapshot, but I'm not sure what to do next to repair the database. This will certainly teach me to make backups of my stuff before major upgrades.

[eliandoran]

@Revali25 , it could be that the TriliumDroid database is slightly different. @FliegendeWurst , could you confirm?

Did you try the other database you mentioned, did you have the same issue with it? Currently in TriliumDroid do you have access to all your notes?

Regardless, you can try opening the database on your computer using https://sqlitebrowser.org/ and look through the notes table.

[FliegendeWurst]

Yes, the TriliumDroid database is slightly different. Most importantly the entity_changes table is only populated with local changes, not synced changes.