Ansible Role: ClickHouse

An Ansible Role that manages installation and configuration of ClickHouse.

Role Variables

Available variables listed below, along with default values (see defaults/main.yml):

clickhouse_conf_dir: /etc/clickhouse-server

Server configuration directory.

clickhouse_loglevel: information

Server log level.

clickhouse_profiles:
  default:
    max_memory_usage: 10000000000
    use_uncompressed_cache: 0
    load_balancing: "random"

A settings profile is a collection of settings grouped under the same name.

clickhouse_profiles:
  default:
    password: "qwerty"
    profile: "default"
    quota: "default"

The users section of the user.xml configuration file contains user settings.

clickhouse_profiles:
  default:
    intervals:
      - duration: 3600
        queries: 0
        errors: 0
        result_rows: 0
        read_rows: 0
        execution_time: 0

Quota settings is for limiting resource usage over a period of time or track the use of resources.

clickhouse_databases:
  - these_metrics
  - other_metrics
  - imported_data

clickhouse_databases is a list of the databases you want to create in ClickHouse.

clickhouse_zookeeper_hosts:
  - 1.2.3.4
  - 1.2.3.5

clickhouse_zookeeper_hosts is the list of zookeeper nodes ClickHouse will connect to. You will need this for replication.

clickhouse_zookeeper_port: 2181

clickhouse_zookeeper_port is the port of zookeeper nodes ClickHouse will connect to. You will need this for replication.

For debugging purposes it is better to start just with one host and then add more hosts.

clickhouse_zookeeper_credentials in format username:password is used to authenticate against Zookeeper.

clickhouse_path: Path to the directory where ClickHouse will store data

clickhouse_log_directory: Path to the error and normal logs.

clickhouse_clusters: is the list of clusters and also contains the shards that are part of each cluster.

clickhouse_clusters:
  distributed_cluster_1:
    shard1:
      - { host: "host_1", port: 9000 }
      - { host: "host_2", port: 9000 }
    shard2:
      - { host: "host_3", port: 9000 }
      - { host: "host_4", port: 9000 }
  distributed_cluster_2:
    shard1:
      - { host: "host_5", port: 9000 }
      - { host: "host_6", port: 9000 }
    shard2:
      - { host: "host_7", port: 9000 }
      - { host: "host_8", port: 9000 }

For the file macros.xml, you need to fill the variables below:

clickhouse_cluster: is the cluster name which is used for replication.

clickhouse_distributed_cluster: is the distributed cluster name defined in clickhouse_clusters.

clickhouse_shard: is the shard number (01, 02) that being part of the replication inside one shard.

Dependencies

None.

Example Playbook

- hosts: all
  roles:
    - nl2go.clickhouse

Development

Use docker-molecule following the instructions to run Molecule or install Molecule locally (not recommended, version conflicts might appear).

Provide Hetzner Cloud token:

export HCLOUD_TOKEN=123abc456efg

Use following to run tests:

molecule test --all

Maintainers

• pablo2go
• dirkaholic

License

See the LICENSE.md file for details.

Author Information

This role was created in 2020 by Sendinblue GmbH.