How to keep the configuration data persistent?

EMQX
1

hi,

I am completely new in the subject of EMQX but I am trying :wink:
I installed EMQX as docker compose but I am missing something - the configuration only is ok until I make command: docker compose rm emqx
But all volumes are on my persistent storage? So why?
here is my compose file:

  emqx:
    image: emqx/emqx:latest
    container_name: emqx
    restart: unless-stopped
    ports:
      - 1883:1883
      - 8083:8083
      - 8084:8084
      - 8883:8883
      - 18083:18083
    volumes:
      - /srv/docker/emqx/data:/opt/emqx/data
      - /srv/docker/emqx/log:/opt/emqx/log
      - /srv/docker/emqx/etc:/opt/emqx/etc

something I am not doing properly - I would appreciate some help :wink:
thx

2

Hi, @zyghom. I’m not sure whether I fully understand what you mean. I think you may be unable to start EMQX because there is no configuration file in /srv/docker/emqx/etc. This is because the container will replace its original directory with the contents of the host directory after mounting.

So you need to start the EMQX container in a non-persistent way first, and then use docker cp command to copy /opt/emqx/etc/emqx.conf to your host machine.

3

thank you @Maverick , this I did already by using this helper - full compose file looks like this:


  init-emqx:
    image: emqx/emqx:latest
    container_name: init-emqx
    entrypoint: sh -c
    volumes:
      # We mount this in the container as `/emqx_etc` to not conflict with built-in files
      # at `/opt/emqx/emqx_etc`
      - /srv/docker/emqx/etc:/emqx_etc
    command:
      # Seed the files from the image to our /emqx_etc if marker file not present, skip
      # if the marker file exists (and implicitly exit cleanly with status code 0)
      - |
        if [ ! -f /emqx_etc/.emqx-seed-complete ]; then
          mkdir /emqx_etc ;
          touch /emqx_etc/.emqx-seed-complete ;
          cp -dpR /opt/emqx/etc/* /emqx_etc/ ;
        fi

  emqx:
    image: emqx/emqx:latest
    container_name: emqx
    restart: unless-stopped
    depends_on:
      init-emqx:
        condition: service_completed_successfully
    
    ports:
      - 1883:1883
      - 8083:8083
      - 8084:8084
      - 8883:8883
      - 18083:18083
    volumes:
      - /srv/docker/emqx/data:/opt/emqx/data
      - /srv/docker/emqx/log:/opt/emqx/log
      - /srv/docker/emqx/etc:/opt/emqx/etc

but my problem is: when I remove container by “docker compose rm emqx” I expect my next start to be with my config: users, passwords etc.
but it is not - why?
I don’t have such problems with other apps: influxdb, grafana, even mosquitto

4

This is because the IP of the container has changed after restarting, so its storage path has also changed.

We list solutions in FAQ, you can look at this question: Why does restarting the EMQX Docker container cause data loss, such as configured rules and resources?

5

perfect - thank you
could you however translate this docker command line to the entries in docker compose file please?
I am not sure how to tackle it
if you could of course :wink:

thank you

6

Of Course, note the environment and networks parts below:

services:
  emqx:
    image: emqx:5.1.3
    container_name: emqx
    environment:
    - "EMQX_HOST={Your Network Alias}"
    networks:
      {Your Network Name}:
        aliases:
        - {Your Network Alias}
7

that will do?

environment:
    - "EMQX_HOST=emqx1.myhost.com"
    networks:
      mynetwork1:
        aliases:
        - emqx1.myhost.com
8

problem solved (and tested)
thank you @Maverick