Database reinitializes every time you run docker compose down then up

[biker2000on]

Symptoms:

1. run openMaint using the default setup with docker-compose up -d
2. Add something to the database
3. docker-compose down
4. docker-compose up -d

In the logs you can see that after step 4 the docker-entrypoint.sh runs the dbcreate command again causing the database to be wiped every time it is started anew. It looks like in docker-entrypoint.sh there used to be some kind of check, but the code will never the section after the ||, it always runs the first command below.

echo "Init DB"
{ # try
  
    $CATALINA_HOME/webapps/cmdbuild/cmdbuild.sh dbconfig create $CMDBUILD_DUMP -configfile $CATALINA_HOME/conf/cmdbuild/database.conf
   
} || { 
    echo "DB was initiliazed. Use dbconfig recreate or dbconfig drop"
}

[biker2000on]

After troubleshooting this issue, it is actually a problem with the database config in docker-compose.yml.

The volume line refers to:

volumes:
  - openmaint-db:/var/lib/postgresql

It should be:

volumes:
  - openmaint-db:/var/lib/postgresql/data

The Dockerfile for the database image calls out Volume: /var/lib/postgresql/data so docker-compose was creating a second temporary volume for the image every time it was stopped and started resulting in all data lost every time.

[afcarvalho1991]

could state the specific docker-compose.yml that you are working on? perhaps using the git URL to that file

[biker2000on]

This was explicitly tested for this line.

https://github.com/itmicus/cmdbuild_docker/blob/master/3.2.1/docker-compose.yml#L11

However, I think this likely applies to all of the docker-compose files because the issue came from docker spinning up a second non-persistent volume for the data directory every time.

[afcarvalho1991]

I was able to reproduce the problem!

When docker-compose down is executed all data stored inside the database is lost.
To reproduce the problem:

docker-compose up -d
# Add something to the database
docker-compose down
docker-compose up -d
# verify that all added data has been lost

Solution:

For all docker-compose.yml files, where:

   volumes:
      - openmaint-db:/var/lib/postgresql

replace by:

    volumes:
      - openmaint-db:/var/lib/postgresql/data

Possible conflicts:

All previous deployments using this new volume structure will be broken, since the directory path will be replaced.

@itmicus what do you suggest we do?

Best Regards,
André.

[afcarvalho1991]

Might update this over the weekend!

[mburger81]

@afcarvalho1991 are there updates for this?

I'm not sure, should we replace the volume mount for that or not? What did you updated as wrote before?

[afcarvalho1991]

As this repo is not mine, it owned by @itmicus I don't feel that I should be making structural changes such as this one without his approval.

[febriansasi]

As this repo is not mine, it owned by @itmicus I don't feel that I should be making structural changes such as this one without his approval.

Please pin it so everyone can read it.
I have same problem and solved it after read this.

[Zakspeed]

OpenMaint data persistence issue openmaint-2.3-3.4.1-d (Docker-compose)

I found a bug in docker-compose that did not allow data to persist when performing a "docker compose down"

SOLUTION:

For all docker-compose.yml files, where:

volumes:

• openmaint-db:/var/lib/postgresql
  replace by:

volumes:

• openmaint-db:/var/lib/postgresql/data

After modifying the line the data is saved perfectly.

IT SEEMS THAT EVERYTHING IS OK

After modifying the line, the data is saved perfectly.

IT SEEMS THAT EVERYTHING IS OK