Setting up using Docker

Troubleshooting
78

OK, so another 40 private replies later, here’s what we also did so everyone still following us here can also participate in the happy end :slight_smile:

After all, the problem was that the nginx proxy manager docker container was in a different docker network scope than the mattermost docker containers and so those two container groups could not communicate with each other.

After reverting all the ports to their defaults and removing the ports: configuration for the mattermost_nginx container in favor of an expose: section, we also found out that the container names were wrong and the nginx proxy manager was referring to a container name that was not available. So we made sure we also aligned these configurations.

Using the command docker network ls showed us a list of multiple different bridge networks and to find out which networks the containers use, we used the following command, which will iterate over all running docker containers and extract the network-related configuration from it and will display it in JSON format:

docker ps --format "{{.ID}}" | xargs docker inspect | jq .[].NetworkSettings.Networks

Here’s an example of what the output looked like:

# mattermost_nginx container
{
  "mattermost": {
    "IPAMConfig": null,
    "Links": null,
    "Aliases": [
      "nginx_mattermost",
      "nginx",
      "d338db70daac"
    ],
[...]

# nginx proxy manager
{
  "docker_default": {
    "IPAMConfig": null,
    "Links": null,
    "Aliases": [
      "nginx_proxy_manager",
      "nginx_proxy_manager",
      "4cf5e1fda301"
    ],
[...]

Here you can see, that the nginx proxy manager is in the network docker_default, whereas the nginx from Mattermost is in the network mattermost.

To fix that, stop the containers and at the bottom of the docker-compose.nginx.yml file, change the networks option to match the desired target network (in this case we moved the Mattermost container group to the docker_default network):

networks:
  default:
    name: docker_default

After we started the containers, we were immediately able to establish the connection and @JPzone282 was up and running after a long journey :slight_smile:

I’m marking this thread as resolved now, I think it’s long enough already.

1 Like