cans.cerebro-container-setup

Provisions and starts a container running Cerebro for ElasticSearch

Requirements

This role has the requirements of Ansible's docker_container module: it requires the docker python module.

Note: Ansible error messages often mention the docker-py module which is deprecated and obsolete. Do install the docker one (use the provided requirements.txt file). If you use specific version of Ansible, check the documentation to install the proper version.

Role Variables

All variables in this roles are namespaced with the cerebrosetup_ prefix. Additionnally variables which are relevant to both the container or the host are prefixed accordingly: cerebrosetup_container_ or cerebrosetup_host_.

Input variables

• cerebrosetup_elasticsearch_hosts: the list of elasticsearch clusters cerebro should connect to. Each element in the list should be as follows:

  uri: "https://cluster.elastic.search:9300/"  # The URI of the cluster
  name: "production cluster"                   # The name to be given to the cluster
  auth:                                        # Credentials to connect to the cluser (optional)
    username: "johndoe"
    password: "some password"
  

  See also the example playbook below. (default: [])

Output variables

• cerebrosetup_container: the role will register the data returned by the module responsible for creating the container in this variable. The content of that variable is described in Ansible's docker_container module documentation

Defaults

• cereborsetup_application_secret: a pseudo random string use to produce cookie signing key;
• cerebrosetup_container_config_dir: the path to cerebro's configuration directory within the container. See also the cerebrosetup_host_config_dir variable. Mostly depends on the images version (default: `"/opt/cerebro/conf");
• cerebrosetup_container_labels: labels to assign to the container (default: {});
• cerebrosetup_container_memory: Maximum amount of memory the container will be allowed to use (default: "512M");
• cerebrosetup_container_name: The name the the container to start, or to be given if provisioning the container is needed (default: "cerebro");
• cerebrosetup_container_network_mode: the Docker network mode the container should user (default: "default")
• cerebrosetup_container_port: TCP port number on which Cerebro will listen to, inside the container (default: 9000);
• cereborsetup_container_rootfs_ro: whether to make the container root filesystem read only (default: false); MUST BE FALSE FOR NOW (looks like cerebro writes a db somewhere, and making the system RO prevents its creation [for http sessions ?]).
• cerebrosetup_container_state: the state the container should end-up in (default: started);
• cerebrosetup_host_bind_address: IP address on which bind the cerebro server running (default: "127.0.0.1");
• cerebrosetup_host_config_dir: directory in which find the application.conf file on the host. It will be mounted within the container (default: "config/");
• cerebrosetup_host_log_dir: where to store logs on the docker host machine. Useful only if you mount want to mount a volume to store logs (default: "/tmp");
• cerebrosetup_host_port: TCP port number on which Cerebro will be published, on the host (default: 9000);
• cerebrosetup_image: name of the container image to use to create the container (default: "yannart/cerebro:latest");
• cerebrosetup_image_pull: whether to force checking if the image is up-to-date, eventually pulling it and updating the container if needed. (default: false);

Dependencies

This role has no dependency.

Example Playbook

This create a container that will run a Cerebro server, reachable on localhost:9000 and that can interact and display data about two Elasticsearch clusters.

- hosts: servers
  roles:
    - role: "cans.cerebro-setup"
      cerebrosetup_container_name: "cerebro-test"
      cerebrosetup_host_config_dir: "~/cerebro/conf"
      cerebrosetup_host_log_dir: "/tmp/logs"
      # The two clusters Cerebro will be capable of interecting with.
      cerebrosetup_elasticsearch_hosts:
        - uri: https://elastic-cluster.example.com:9300/
          name: "The Elastic Cluster"
          auth:
            username: "bar"
            password: "bazword"
        - uri: https://unyielding-cluster.example.com:9300/
          name: "The Unyielding Cluster"
          # If no credentials are required, omit the `auth` key.

TODO

• Manage to generate an application secret (not so important if used bound to the loopback address, as is by default);

License

GPL 2.0

Author Information

Copyright © 2018, Nicolas CANIART.