Backups and restore

Everything Day Planner stores (accounts, tasks, documents, settings) is in its Postgres database. Back that up and you can rebuild the whole instance.

Make a backup

This saves a compressed dump named with the current time.

docker compose exec -T db pg_dump -U dayplanner -d day_planner -Fc \
  > day_planner-$(date -u +%Y%m%dT%H%M%SZ).dump

Check the file starts with PGDMP, the custom-format header:

head -c 5 day_planner-*.dump

Copy backups off the server. Run this from cron for regular backups.

Test a backup

A backup that’s never been restored isn’t a backup. This restores into a throwaway container and counts users:

docker run -d --name dp-restore -e POSTGRES_PASSWORD=restore -e POSTGRES_DB=day_planner postgres:16-alpine
sleep 5
docker cp day_planner-<timestamp>.dump dp-restore:/tmp/b.dump
docker exec dp-restore pg_restore -U postgres -d day_planner --no-owner --no-acl /tmp/b.dump
docker exec dp-restore psql -U postgres -d day_planner -Atc "select count(*) from users"
docker rm -f dp-restore

Compare the count, and those for tasks and documents, with the live database.

Restore

Restoring replaces the live data. Take a fresh backup first, and stop the app so nothing writes during the restore.

docker compose stop app
docker compose cp day_planner-<timestamp>.dump db:/tmp/b.dump
docker compose exec db pg_restore -U dayplanner -d day_planner --clean --if-exists --no-owner --no-acl /tmp/b.dump
docker compose start app

When the app starts, it applies any migrations the backup is missing.

See also

• Upgrading: back up before every upgrade
• Migration failed: recovering from a failed upgrade