Map a Docker app’s data before an upgrade
A short, read-only inventory shows which files survive a container replacement and which still need a backup plan.

In this article
Before you start
You need access to the Docker CLI on the host and the name of the running container. This documentation-based walkthrough inspects metadata. It has not been run against your NAS.
1. Find the mounts
Run docker ps to identify the container. Then run docker inspect --format '{{json .Mounts}}' CONTAINER, substituting its name. Record each mount’s type, source and destination. For a named volume, docker volume inspect VOLUME shows the volume metadata.
2. Match paths to the application
Use the app’s documentation to map database files, uploaded media and configuration to those destinations. A file written only into the container’s writable layer can disappear when the container is replaced. A named volume can survive replacement while sharing the same physical disk as the running service.
3. Check recovery coverage
Match every irreplaceable path to a backup job and its most recent successful run. For a database, follow the application’s consistent-backup procedure. Validate recovery in a separate location. The inventory itself makes no changes and requires no rollback.
Follow the file to the host
Pick one settings file and one data file your app reads. In the mount inventory, identify the destination path inside the container and the matching source. A bind mount should point to a known path. A named volume should have an identifiable volume name. If a file lives only in the container's writable layer, replacing the container can remove the file from the running application. Record the ownership and permissions the app expects. A path can exist and still fail if the process cannot read it.
Draw a small inventory with columns for app path, persistent source, backup method and recovery check. For a database, record its supported dump or consistent snapshot method. A directory copy taken while data is changing gives you different guarantees than a restore format the database supports.
Rehearse an upgrade with a separate copy
Restore into a temporary instance with a different project name and ports. Verify that it is using the copies of the mounts before starting it. Check a known setting and one record through the application. Then take the upgrade on the test instance. If the upgrade changes the database schema, document recovery from the pre-upgrade backup before relying on the previous image tag as a rollback.
Primary sources
Docker volume documentationRead next Local or cloud? Follow the work.