Reading time: 2 minutes.

This section gives information about CAM installation with a CAM Docker image.

CAM Docker configuration

CAM environment variables

CAM mount points

Python ML libs

[3.10] Python ML libs are installed inside CAM Docker image so there is no need for unzipping a ML.zip containing these libraries. spaCy models are not contained into these installed libraries anymore and are autoloaded in the CAM configuration.

For more information please have a look at CAM Python ML libraries documentation.

CAM Docker installation

Docker installation

Install Docker and Docker Compose accordingly to your OS instructions: https://docs.docker.com/install/

Create CAM Docker image

1. Unzip CAM Docker package docker-CAM-${CAM_VERSION}-image.zip and go inside the root directory.
2. Run the command:

sudo docker image build -t mondeca/cam:${CAM_VERSION} .

Deploy CAM Docker image

1. Install needed dependencies like KC, Postgresql, ITM (you can skip this step if you run without authentication server and without external database using internal H2 DB)
2. Unzip default CAM conf in folder content-augmentation
3. Create logs folder
4. Create work folder

With docker

1. Create container with command: 

sudo docker create --name cam -p 8180:8180 --mount type=bind,source="$(pwd)"/content-augmentation,target=/opt/jboss/wildfly/cam/configuration/content-augmentation --mount type=bind,source="$(pwd)"/log,target=/opt/jboss/wildfly/cam/log --mount type=bind,source="$(pwd)"/py-envs,target=/opt/jboss/wildfly/cam/bin/py-envs  -it mondeca/cam:{CAM_VERSION}

2. Run container with command:

sudo docker start cam

3. Stop container with command:

sudo docker stop cam

With docker-compose

1. Update docker-compose.yaml with appropriate values

2. Start CAM container with command:

sudo docker compose up -d

3. Stop container with command:

sudo docker compose down

Going inside the container

sudo docker exec -it docker-cam-1 bash

Removing the container

sudo docker rm cam

Running CAM in HA

There is a docker-compose-ha-local-test to test CAM HA with 2 nodes on localhost.

If you want to run several HA nodes with a CAM Docker image:

- update docker-compose.yaml with appropriate values

- STANDALONE_CONF must be set to standalone-ha.xml

- Using a Postgres DB is mandatory

- Docker network must be set so CAM containers can communicate on private network through additional ports (7600; 7650)

CAM with GPU

Please follow these installation steps on the host:

curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

sudo apt-get update

sudo apt-get install -y nvidia-container-toolkit

sudo nvidia-ctk runtime configure --runtime=docker

sudo service docker restart

To create CAM GPU Docker image: use the docker file: Dockerfile_gpu.

To run with Docker: 

sudo docker run --name cam --gpus all --runtime:nvidia mondeca/cam:${CAM_VERSION}

To run with Docker Compose: uncomment in docker-compose.yaml the GPU part (environment variables, runtime).