Clickhouse Cluster On-Prem Deployment

This docker compose is to enable Clickhouse on-prem deployment with Zookeeper and Multi-Replica Clickhouse with storage as Wasabi/S3 buckets

Why Wasabi?

Wasabi is extremely low-cost compared to AWS S3 and does not charge any egress fee!

How does this work?

1. Zookeeper syncs data between multiple clickhouse replicas
2. Clickhouse by defaults uses local disk to store the data and every 10s sync to Wasabi/S3.

Tip

You can also update the storage config tiered for s3 as "cold" and use a TTL table in clickhouse to sync every few seconds.

How to run?

docker compose up -d

Configurations

1. Change your Bucket endpoint inside the clickhouse-storage.xml

    <endpoint>https://s3.wasabisys.com/WASABI_BUCKET_NAME/data/</endpoint>
    <access_key_id>WASABI_ACCESS_KEY_ID</access_key_id>
    <secret_access_key>WASABI_SECRET_ACCESS_KEY</secret_access_key>

2. Set the Username and Password for Clickhouse in docker-compose.yml

    environment:
      - CLICKHOUSE_USER=<USERNAME>
      - CLICKHOUSE_PASSWORD=<PASSWORD>

3. Add multi-replicas/shards in cluster-config.xml

    <shard>
        <replica>
            <host>clickhouse-2</host>
            <port>9000</port>
        </replica>
    </shard>