Unable to connect to container's browser UI.

[Com177]

Ask a Question!

I feel stupid but I can't figure out what I'm doing wrong. I just deployed my first container in Portainer as a Stack. The container is NetAlertX. The stack is healthy and I mapped the container's recommended host port, 20211, to container port 8080 in the compose file. I also added the IP address of "my server" to the "local" Environment where the container is. When I try to launch the NetAlertX UI in my browser it will not connect. So, I need to know if I am making a bad assumption when it comes to the IP of my server.

My setup is as follows; I have a Zimaboard on my network that I did a bare-metal install of Proxmox VE v9.0. I installed the Docker LXC container into Proxmox and during the install I also opted to install Portainer. My Portainer container has a network IP of 192.168.1.10 and Proxmox VE has a network IP of 192.168.1.248. I thought I should be using the Proxmox address with the NetAlertX port (https://192.168.1.248:20211) but I get nothing. I tried using http:// but I get the same result. The same thing happens if I use the Portainer IP. Should I be using the Proxmox IP or the Portainer IP to get into the container?

Here is my compose file if it helps. I could really use some advice. Thanks.

services:
  netalertx:
     image: jokobsk/netalertx:latest
     container_name: netalertx
     #healthcheck:
      #test: curl -f http://192.168.1.248:20211/ || exit 1
     #mem_limit: 4g
     #security_opt:
      #- no-new-privileges:true
     volumes:
      - config:/app/config:rw
      - db:/app/db:rw
      - logs:/app/front/log:rw
     ports:
      - 20211:8080
     environment:
      TZ: America/New_York
      PORT: 20211
      HOST_USER_ID: 1000
      HOST_USER_GID: 1000
      network_mode: host
      restart: on-failure:5
     
volumes:
  config:
  db:
  logs:

[Nick-Portainer]

Hey @Com177

Use the IP of the Docker host where NetAlertX is running. In your case that is the IP of the LXC container that’s running Docker, not the Proxmox node IP and not the Portainer container’s IP. Find it inside the LXC with hostname -I and then browse to http://'LXC-IP':20211.

You also have a couple of compose issues that will block access:

1. Port mapping is wrong
   NetAlertX listens on 20211 in the container. You mapped 20211:8080, which points to a port the app is not using. Use 20211:20211 instead, or change the app to listen on 8080 and keep the mapping consistent.

2. network_mode and restart are in the wrong place
   They are under environment right now, so they are being treated as environment variables rather than compose keys.

Here are is a correct compose option.

Option A: default bridge networking with a published port

services:
  netalertx:
    image: jokobsk/netalertx:latest
    container_name: netalertx
    restart: on-failure:5
    volumes:
      - config:/app/config:rw
      - db:/app/db:rw
      - logs:/app/front/log:rw
    ports:
      - "20211:20211"      # host:container
    environment:
      TZ: America/New_York
      HOST_USER_ID: 1000
      HOST_USER_GID: 1000
      # PORT is optional if you stick with 20211 default
      # PORT: 20211

volumes:
  config:
  db:
  logs:

[Com177]

Thank you @Nick-Portainer !!!

I didn't know that my compose file was incorrect and I plan to redeploy the container using your fixes. I see the revised compose file no longer has "network mode" in it. Is it OK not to have it? My container hostname is docker-portainer.

I tried to use -I and I got this

If I ping the container hostname I get this:

When I exit the console I see this:

It seems I can't separate the Portainer UI from the docker container itself. Should they be separable? I will implement your fixes and try the external IP's that I know are real (Portainer & Proxmox) and see what happens. Let me get back to you later today with the results.

Thanks so much for the reply!

[Com177]

[SOLVED]- Hi @Nick-Portainer !!!! THANK YOU!! THANK YOU!! THANK YOU!! Your corrections to the compose file did the trick!!! The container deployed with only one error but it looks like it isn't a critical error.

From the advice you gave about the IP, I took the hint from all of the clues I found pointing to the container's external IP being 192.168.1.10 . I searched the internet and found that if I open the Console in the container itself and type ip a I got a list of all of the interfaces for the container, much like ifconfig does in SSH. If I go to the proxmox "node" that contains the container and open a Shell and type lxc-info -n <container ID> I get all of the IP addresses that the container is listening on. These two ways all pointed to the Portainer external IP; in my case 192.168.1.10 being the IP to use for the UI. Going with the flow I was able to open the NetAlertX UI using the Portainer IP and the NetAlertX port. So, 192.168.1.10:9443 brings me to Portainer and 192.168.1.10:20211 brings me to NetAlertX. I also first had to set the Public IP of my Environment to get it to work (found under: Environment > Edit > Configuration). It was blank so I entered the Portainer IP.

I am seeing a new problem now but I want to try to find a solution on my own before I ask you or anyone else for advice. Briefly, NetAlertX uses ARPSCAN to find devices on a LAN. It looks like other than my WAN gateway and the internal IP for the container itself (172.18.0.2), NetAlertX cannot find anything else. I believe the problem is that I need to create a bridge of some kind between Docker/Portainer and Proxmox to allow NetAlertX to arpscan my entire LAN. I have Pi-Hole running in a container directly on Proxmox and it can find all of my devices, so I think I need to find out how to make a useable interface between Docker/Portainer and Proxmox so that NetAlertX can scan my network.

Thanks so much for your help. It made it possible for me to get NetAlertX up and running! Thanks again!!