How to restore the zammad backup?

[macnibblet]

I cannot find any documentation on how to restore a backup made

[monotek]

Because there is none...

Just copy files and import DB manually in the containers.

[macnibblet]

I have spent a few hours trying to restore from a none docker-compose backup without any luck. Any chance you could walk me through it?

[lbausch]

@macnibblet
Make sure you have your files and your SQL dump ready.

Files should go into /var/lib/docker/volumes/zammaddockercompose_zammad-data/_data. You can check if this is the correct path by running docker inspect zammaddockercompose_zammad-railsserver_1 and see the "Mounts" section which might look like this:

"Mounts": [
            {
                "Type": "volume",
                "Name": "zammaddockercompose_zammad-data",
                "Source": "/var/lib/docker/volumes/zammaddockercompose_zammad-data/_data",
                "Destination": "/opt/zammad",
                "Driver": "local",
                "Mode": "rw",
                "RW": true,
                "Propagation": ""
            }
        ],

Copy your files there.

For SQL dump: Make sure to stop all containers except the postgresql container by running docker stop <container name>.
To access the SQL dump inside the container you need to check which mounts the container has:
docker inspect zammaddockercompose_zammad-postgresql_1 (you containers might be named differently).
In my case the volume path mounted in the container is /var/lib/docker/volumes/zammaddockercompose_postgresql-data/_data. Copy your backup there.

Next launch a shell inside the postgresql container: docker exec -it zammaddockercompose_zammad-init_1 bash. Switch to the postgres user by running su - postgres. Change the directory to /var/lib/postgresql/data and you should see your backup there. If it is a .gz file extract it using gunzip.
To get a clean restore you need to drop the existing database by running dropbdb zammad_production. Recreate it with createdb zammad_production. Finally, import the dump with psql zammad_production < dump.sql.
Once you're done exit the container, stop it and use docker-compose to start all Zammad containers.

Disclaimer: I'm by no means an Postgres guru so YMMV. :)

[macnibblet]

@lbausch This worked almost flawlessly! The only problems is that the files keep getting written over each time I start it and we had replaced the OAuth.rb with a custom solution for hub from JetBrains. But I moved the account login to google OAuth instead. Thanks!

This should definitely be added to a guide somewhere or turned into a bash script

[waja]

I did it this way:

• Needed volumes in the Postgresql container:

  zammad-postgresql:
    volumes:
      - zammad-data:/opt/zammad
      - zammad-backup:/var/tmp/zammad:ro

• Stopping services

# docker-compose stop zammad-nginx zammad-websocket zammad-railsserver zammad-memcached zammad-elasticsearch zammad-scheduler

• Drop database content

# docker-compose exec zammad-postgresql su -c 'dropdb zammad_production' - postgres

• Start Scheduler again (to recreate empty db)

# docker-compose start zammad-scheduler

• Recreate empty database

# docker-compose exec zammad-scheduler bundle exec rake db:create

• (Preparing) Restore

# docker-compose exec zammad-postgresql /bin/sh
$ cp /opt/zammad/contrib/backup/config.dist /opt/zammad/contrib/backup/config && \
  sed -i s/zammad_backup/zammad/g /opt/zammad/contrib/backup/config && \
  ln -s /usr/local/bin/psql /usr/bin/ && \
  /opt/zammad/contrib/backup/zammad_restore.sh

• Restart whole compose stack

# docker-compose stop; docker-compose start && docker-compose logs -f

[monotek]

Currently we're not planning to script this in any way.
Feel free to add a pull request if you have a working solution.
Closing for now.

[apiening]

# docker-compose exec zammad-postgresql su -c 'dropdb zammad_production' - postgres

I get the the following error when I try to execute this command:

dropdb: could not connect to database template1: FATAL:  role "postgres" does not exist

Any idea why this is?

[MrGeneration]

Please don't recycle long solved issues for such things.
Either open a new issue on this repo or (better) use the community at https://community.zammad.org .

[linucksrox]

@MrGeneration Sorry but this is not a long solved issue. The command doesn't work, presumably because something changed in how the database container runs since the original solution was posted.

Why shut down the conversation here in favor of opening new threads? That just makes it more difficult to find an updated solution. Even if you don't support Docker, there's no need to artificially make it more difficult for the community to collaborate with each other.

I'd rather have one thread with all the information instead of multiple dead ends on the community forum as it is today.

[apiening]

@linucksrox I ended up using https://github.com/loomchild/volume-backup to backup and restore my zammad docker volumes.
I'm aware that this is not an optimal solution, but I was not able to do a successful restore based on the mechanism the zammad documentation described at this time.

[linucksrox]

Thanks @apiening! I'm currently checking out https://github.com/DoTheEvo/selfhosted-apps-docker/tree/master/zammad#backup-and-restore but if that doesn't work for me then I'll try your solution.

[mmojadad]

@waja
Hi
First thank for your good guides.
Question : I am using Ubuntu 20.04
Issue is as following

docker-compose exec zammad-postgresql /bin/sh $ cp /opt/zammad/contrib/backup/config.dist /opt/zammad/contrib/backup/config && \ sed -i s/zammad_backup/zammad/g /opt/zammad/contrib/backup/config && \ n -s /usr/local/bin/psql /usr/bin/ && \ /opt/zammad/contrib/backup/zammad_restore.sh

/bin/sh: can't open '$': No such file or directory