Running MikoPBX using docker compose

To work with MikoPBX in a container, you need to install Docker and Docker Compose following the instructions

Docker installation and creating a user and directories

Docker compose launch option

Here is an example of a docker-compose.yml file that can be used to manage your MikoPBX container via Docker Compose:

docker-compose.yml
services:
  mikopbx:
    container_name: "mikopbx"
    image: "ghcr.io/mikopbx/mikopbx-x86-64"
    network_mode: "host"
    cap_add:
      - NET_ADMIN
    entrypoint: "/sbin/docker-entrypoint"
    hostname:  "mikopbx-in-a-docker"
    volumes:
      - /var/spool/mikopbx/cf:/cf
      - /var/spool/mikopbx/storage:/storage
    tty: true
    environment:
      - ID_WWW_USER=${ID_WWW_USER}
      - ID_WWW_GROUP=${ID_WWW_GROUP}
      # Change the station name through environment variables
      - PBX_NAME=MikoPBX-in-Docker
      # Change the default SSH port to 23
      - SSH_PORT=23
      # Change the default WEB port to 8080
      - WEB_PORT=8080
      # Change the default WEB HTTPS port to 8443
      - WEB_HTTPS_PORT=8443

Save the contents into a file named docker-compose.yml, make the necessary adjustments, and launch MikoPBX using the command:

export ID_WWW_USER=$(id -u www-user)
export ID_WWW_GROUP=$(id -g www-user)
sudo docker compose -f docker-compose.yml up

Running Multiple MikoPBX Instances on One Host

Mode Without Network Isolation Between Host and Containers (–net=host)

It is also possible to organize the launch of multiple MikoPBX containers on a single host. However, you need to consider Docker's port handling features. If the –net=host mode is not used, it will lead to a high load on the host system's CPU because Docker creates a separate rule in Iptables for each allocated port.

With the –net=host mode enabled, you need to manually monitor the distribution of available ports between the running containers and built-in applications. For instance, to run two MikoPBX containers on one host, you can use the following configuration file:

docker-compose.yml
services:
  mikopbx-first:
    container_name: "mikopbx-first"
    image: "ghcr.io/mikopbx/mikopbx-x86-64"
    network_mode: "host"
    entrypoint: "/sbin/docker-entrypoint"
    hostname: "mikopbx-in-docker-first"
    volumes:
      - /var/spool/mikopbx/first/cf:/cf
      - /var/spool/mikopbx/first/storage:/storage
    tty: true
    environment:
      - ID_WWW_USER=${ID_WWW_USER}
      - ID_WWW_GROUP=${ID_WWW_GROUP}
      - PBX_NAME=MikoPBXFirst
      - PBX_FIREWALL_ENABLED=0
      - PBX_FAIL2BAN_ENABLED=0
      - SSH_PORT=123
      - WEB_PORT=8080
      - WEB_HTTPS_PORT=8443
      - SIP_PORT=5060
      - TLS_PORT=5061
      - RTP_PORT_FROM=10000
      - RTP_PORT_TO=10800
      - IAX_PORT=4569
      - AMI_PORT=5038
      - AJAM_PORT=8088
      - AJAM_PORT_TLS=8089
      - BEANSTALK_PORT=4229
      - REDIS_PORT=6379
      - GNATS_PORT=4223
  mikopbx-second:
    container_name: "mikopbx-second"
    image: "ghcr.io/mikopbx/mikopbx-x86-64"
    network_mode: "host"
    tty: true
    entrypoint: "/sbin/docker-entrypoint"
    hostname: "mikopbx-in-docker-second"
    volumes:
      - /var/spool/mikopbx/second/cf:/cf
      - /var/spool/mikopbx/second/storage:/storage
    environment:
      - ID_WWW_USER=${ID_WWW_USER}
      - ID_WWW_GROUP=${ID_WWW_GROUP}
      - PBX_NAME=MikoPBXSecond
      - PBX_FIREWALL_ENABLED=0
      - PBX_FAIL2BAN_ENABLED=0
      - SSH_PORT=2223
      - WEB_PORT=8081
      - WEB_HTTPS_PORT=9443
      - SIP_PORT=6060
      - TLS_PORT=6061
      - RTP_PORT_FROM=20000
      - RTP_PORT_TO=20800
      - IAX_PORT=5569
      - AMI_PORT=6038
      - AJAM_PORT=9088
      - AJAM_PORT_TLS=9089
      - BEANSTALK_PORT=5229
      - REDIS_PORT=7379
      - GNATS_PORT=5223

Save the contents into a file named docker-compose.yml, make the necessary adjustments, and launch MikoPBX using the command:

export ID_WWW_USER=$(id -u www-user)
export ID_WWW_GROUP=$(id -g www-user)
sudo docker compose -f docker-compose.yml up

Network Bridge Mode (–net=bridge)

There is an option to launch MikoPBX containers in the –net=bridge mode. However, as mentioned above, to use this mode you either need to significantly limit the range of RTP ports or open them on the host machine without using Docker's capabilities.

For this, you will need to write a small script to determine the name of the current bridge interface and the IP address of each container. After running Docker Compose, you will then need to add the necessary iptables rules for the RTP port range as follows:

start-multiple-mikopbx.sh
#!/bin/bash

COMPOSE_FILE="$1"

if [ -z "$COMPOSE_FILE" ]; then
    echo "Usage: $0 path/to/docker-compose.yaml"
    exit 1
fi

# We will obtain the user ID for running the container
export ID_WWW_USER=$(id -u www-user)
export ID_WWW_GROUP=$(id -g www-user)

# Stop current containers if they are running
docker compose -f "$COMPOSE_FILE" down

# Remove them
docker compose -f "$COMPOSE_FILE" rm

# Start containers in the background
docker compose -f "$COMPOSE_FILE" up -d
sleep 60

# Create a label for IPTABLES rules
IPTABLES_COMMENT="mikopbx-custom-rule"

# Determine the project identifier, used when creating a network bridge
project_prefix=$(cat "$COMPOSE_FILE" | yq e '.x-project-name' -)

# If the prefix is not set, use a default value
if [ -z "$project_prefix" ]; then
    project_prefix="default_prefix"
fi

# Function to get container IP address
function get_container_ip() {
    docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$1"
}

# Function to get the name of the bridge interface
function get_bridge_name() {
    local network_name="$1"
    local prefix="$2"
    local network_id=$(docker network inspect "${prefix}_${network_name}" -f '{{.Id}}')

    if [ -z "$network_id" ]; then
        echo "Error: Network ${prefix}_${network_name} not found."
        return 1
    fi

    local bridge_name=$(ip link show type bridge | grep -o "br-${network_id:0:12}\b")
    echo $bridge_name
}

echo "Delete tagged iptables rules"
# Delete all iptables rules tagged with our comment
iptables -S | grep "$IPTABLES_COMMENT" | sed 's/-A /-D /' | while read rule; do
        echo "Delete rule $rule"
        iptables $rule
done

# Delete all NAT iptables rules tagged with our comment
iptables -S -t nat | grep "$IPTABLES_COMMENT" | sed 's/-A /-D /' | while read rule; do
     echo "Delete rule $rule"
     iptables -t nat $rule
done

# Parse the docker-compose file and obtain all necessary parameters.
echo "Parsing docker-compose file and configuring iptables rules"
cat "$COMPOSE_FILE" | yq e '.services[] | select(.environment[] | test("RTP_PORT_FROM")) | {"container_name": .container_name, "environment": .environment, "network": .networks[0]}' -o=json | jq -c '.' | while read -r service; do
    container_name=$(echo $service | jq -r '.container_name')
    network_name=$(echo $service | jq -r '.network')
    bridge_name=$(get_bridge_name "$network_name" "$project_prefix")
    container_ip=$(get_container_ip "$container_name")

    RTP_PORT_FROM=$(echo $service | jq -r '.environment[] | select(contains("RTP_PORT_FROM")) | split("=")[1]')
    RTP_PORT_TO=$(echo $service | jq -r '.environment[] | select(contains("RTP_PORT_TO")) | split("=")[1]')

    echo "Configuring iptables for $container_name ($container_ip) on $bridge_name from port $RTP_PORT_FROM to $RTP_PORT_TO"

    iptables -A DOCKER -t nat ! -i "$bridge_name" -p udp -m udp --dport $RTP_PORT_FROM:$RTP_PORT_TO -j DNAT --to-destination $container_ip:$RTP_PORT_FROM-$RTP_PORT_TO -m comment --comment "$IPTABLES_COMMENT"
    iptables -A DOCKER -d $container_ip/32 ! -i "$bridge_name" -o "$bridge_name" -p udp -m udp --dport $RTP_PORT_FROM:$RTP_PORT_TO -j ACCEPT -m comment --comment "$IPTABLES_COMMENT"
    iptables -A POSTROUTING -t nat -s $container_ip/32 -d $container_ip/32 -p udp -m udp --dport $RTP_PORT_FROM:$RTP_PORT_TO -j MASQUERADE -m comment --comment "$IPTABLES_COMMENT"

    echo "Don't forget to open UDP ports $RTP_PORT_FROM to $RTP_PORT_TO on external firewall if it exists"
done

echo "iptables configuration completed successfully."

Let's describe several containers in the docker-compose.yaml file, specify different ports for the web interface, SIP ports, and ranges of RTP ports to ensure they do not overlap.

docker-compose.yaml
services:
  mikopbx-first:
    container_name: "mikopbx-first"
    image: "ghcr.io/mikopbx/mikopbx-x86-64"
    entrypoint: "/sbin/docker-entrypoint"
    hostname:  "mikopbx-in-docker-first"
    volumes:
      - /var/spool/mikopbx/first/cf:/cf
      - /var/spool/mikopbx/first/storage:/storage
    tty: true
    cap_add:
      - net_admin
    networks:
      - network-bridge1
    environment:
      - ID_WWW_USER=${ID_WWW_USER}
      - ID_WWW_GROUP=${ID_WWW_GROUP}
      - PBX_NAME=MikoPBXFirst
      - RTP_PORT_FROM=10000 # UDP range 10000-10800 on host will be directed to the container
      - RTP_PORT_TO=10800
      - WEB_ADMIN_PASSWORD=mikopbx-first-password
      - ENABLE_USE_NAT=1
      - PBX_FIREWALL_ENABLED=1
      - PBX_FAIL2BAN_ENABLED=1
    ports:
      - "8443:443"  # TCP port 8443 on the host is directed to port 443 in the container
      - "5060:5060/udp"  # UDP port 5060 on the host is directed to port 5060 in the container
  mikopbx-second:
    container_name: "mikopbx-second"
    image: "ghcr.io/mikopbx/mikopbx-x86-64"
    tty: true
    cap_add:
      - net_admin
    networks:
      - network-bridge2
    entrypoint: "/sbin/docker-entrypoint"
    hostname:  "mikopbx-in-docker-second"
    volumes:
      - /var/spool/mikopbx/second/cf:/cf
      - /var/spool/mikopbx/second/storage:/storage
    environment:
      - ID_WWW_USER=${ID_WWW_USER}
      - ID_WWW_GROUP=${ID_WWW_GROUP}
      - PBX_NAME=MikoPBXSecond
      - RTP_PORT_FROM=20000 # UDP range 20000-20800 on host will be directed to the container
      - RTP_PORT_TO=20800
      - EXTERNAL_SIP_PORT=6060 # Inform MikoPBX about its external SIP port
      - WEB_ADMIN_PASSWORD=mikopbx-second-password
      - ENABLE_USE_NAT=1
      - PBX_FIREWALL_ENABLED=1
      - PBX_FAIL2BAN_ENABLED=1
    ports:
      - "9443:443"  # TCP port 9443 on the host is directed to port 443 in the container
      - "6060:5060/udp"  # UDP port 6060 on the host is directed to port 5060 in the container
x-project-name: mikopbx # This parameter must be present
networks:
  network-bridge1:
    driver: bridge
  network-bridge2:
    driver: bridge

Creating a directory for scripts

mkdir -p /usr/src/mikopbx

Save the start-multiple-mikopbx.sh and docker-compose.yaml files into this folder.

Install the necessary dependencies for the script.

sudo apt-get update
sudo apt-get install jq
sudo snap install yq

Navigate to our folder, add execution rights and launch our script.

cd /usr/src/mikopbx
sudo chmod +x start-multiple-mikopbx.sh
sudo ./start-multiple-mikopbx.sh docker-compose.yaml

While waiting for the containers to start, check the firewall settings on the host, and if necessary, open the ports specified in our docker-compose.yaml file, specifically:

  • TCP/UDP ports 5060 and 6060 for SIP

  • UDP ranges 10000-10800 and 20000-20800 for RTP voice transmission

  • TCP ports 8443 and 9443 for HTTPS protocol, for web interface operation.

Access each station in turn at the addresses:

  • https://<host machine IP>:8443

  • https://<host machine IP>:9443

To access the web interface of the first MikoPBX, use the login admin and the password mikopbx-first-password

To access the web interface of the second MikoPBX, use the login admin and the password mikopbx-second-password

Each machine should have NAT mode enabled, indicating that the container is behind a router in the network interface settings. If the stations will be used within a local network, then in the external IP field, enter the local IP address of the host machine, otherwise its public IP address.

Important note! One of our containers uses port forwarding from SIP port changing its value from 5060 to 6060. In this case, for the system to function correctly, you need to add the external value of the SIP port in the NAT settings in the network interfaces section of MikoPBX. This setting can also be made by setting the corresponding value of the environment variable EXTERNAL_SIP_PORT=6060 in the docker-compose file.

With that, the setup is complete, and you can configure accounts and make calls.

Environment variables for configuring MikoPBX

Below are some of the environment variables that will allow you to adjust the MikoPBX ports and settings used.

  • SSH_PORT - port for SSH (22)

  • WEB_PORT - port for the web interface via HTTP protocol (80)

  • WEB_HTTPS_PORT - port for the web interface via HTTPS protocol (443)

  • SIP_PORT - port for connecting a SIP client (5060)

  • TLS_PORT - port for connecting a SIP client with encryption (5061)

  • RTP_PORT_FROM - beginning of the RTP port range, voice transmission (10000)

  • RTP_PORT_TO - end of the RTP port range, voice transmission (10800)

  • IAX_PORT - port for connecting IAX clients (4569)

  • AMI_PORT - AMI port (5038)

  • AJAM_PORT - AJAM port used for connecting the telephony panel for 1C (8088)

  • AJAM_PORT_TLS - AJAM port used for connecting the telephony panel for 1C (8089)

  • BEANSTALK_PORT - port for the Beanstalkd queue server (4229)

  • REDIS_PORT - port for the Redis server (6379)

  • GNATS_PORT - port for the gnatsd server (4223)

  • ID_WWW_USER - identifier for www-user (can be set with the expression $(id -u www-user), where www-user is NOT a root user)

  • ID_WWW_GROUP - group identifier for www-user (can be set with the expression $(id -g www-user), where www-user is NOT a root group)

  • WEB_ADMIN_LOGIN - login for Web interface access

  • WEB_ADMIN_PASSWORD - password for Web interface access

A full list of all possible setting parameters is available in the source code here.

Last updated