Install Hyper-Converged Database (HCD) using Docker for non-production local containerized environments

Set up HCD in a local containerized environment for development and testing purposes. This installation provides a way to explore HCD features without complex infrastructure setup.

Use HCD in Docker for development and testing only. Don’t use this installation method for production deployments. For production environments, use Mission Control or binary tarball installation.

What is included in the HCD development installation?

This installation method is a containerized version of HCD that includes:

  • One HCD node

  • One Data API instance

  • Data API support

  • CQL support

  • Pre-configured Docker Compose setup for local development

Prerequisites

  • Docker Engine 20.10 or later installed and running

  • Docker Compose version 2.0 or later

  • At least 6 GB of available memory for the Docker containers

  • At least 1.6 GB of available disk space for images

  • At least 1 GB of available disk space for data and logs

Authenticate with IBM Container Registry (ICR)

Before you pull the HCD Docker images, authenticate with the ICR:

  1. Navigate to MyIBM.

  2. Click Container Software & Entitlement Keys.

  3. Copy an existing entitlement key or create a new one.

  4. Authenticate with your entitlement key:

    docker login cp.icr.io --username cp

    When prompted for a password, paste your entitlement key.

    You should see:

    Login Succeeded

    If you don’t see this message, check your entitlement key and try again.

Download the configuration files

Download the Docker Compose configuration files to your local machine:

  1. Create a directory for the installation:

    mkdir hcd-dev && cd hcd-dev

  2. Download the compose.yaml file to this directory.

  3. Download the cassandra.yaml configuration file to this directory.

    Both files must be in the same directory where you run docker compose up.

Launch HCD with Docker Compose

Start the HCD node and Data API using Docker Compose:

docker compose up -d

This command does the following:

  • Pulls the HCD and Data API images from IBM Container Registry.

  • Starts the HCD node with a 3 GB memory limit.

  • Starts the Data API on port 8181 with a 2 GB memory limit.

  • Creates health checks to ensure services are ready.

The initial startup might take several minutes as the images download and the database initializes.

Verify the installation

Check that the containers are running and healthy.

  1. View the container status:

    docker compose ps

    You should see both hcd and data-api containers with status "Up" and "healthy".

  2. View the logs to monitor startup progress:

    docker compose logs -f

    Press Ctrl+C to stop following the logs.

Connect to HCD

After the containers are healthy, you can connect to HCD using several methods.

Container names include your directory name as a prefix. If you created the installation in a directory called hcd-dev, the containers are named hcd-dev-hcd-1 and hcd-dev-data-api-1. Adjust the container names in the following commands based on your directory name.

Use CQL shell (cqlsh)

Connect to the running container and start cqlsh:

docker exec -it hcd-dev-hcd-1 cqlsh -u cassandra -p cassandra

You should see the CQL shell prompt:

Connected to hcd-cluster at 192.0.2.200:9042
[cqlsh 6.1.0 | Cassandra 4.1.x | CQL spec 3.4.6 | Native protocol v5]
Use HELP for help.
cqlsh>

The default credentials are username cassandra and password cassandra.

Use the Data API

The Data API is available at http://localhost:8181. You can use the Swagger UI to test the API at http://localhost:8181/swagger-ui/.

Test the connection:

curl http://localhost:8181/stargate/health

For more information about the Data API, see Get started with the Data API.

Use Data API clients or Cassandra drivers

You can connect your applications using either Data API clients or Cassandra drivers, depending on your use case.

Data API clients

For JSON document operations and vector search, use the Data API clients to connect to http://localhost:8181.

For Data API client installation and usage, see Get started with the Data API.

Cassandra drivers

For CQL operations, connect your applications using standard Cassandra drivers on localhost:9042.

Use the following connection details:

  • Host: localhost

  • Port: 9042

  • Username: cassandra

  • Password: cassandra

  • Local datacenter: dc1

For driver installation and usage, see DataStax driver documentation.

View container logs

View HCD logs:

docker compose logs hcd

View Data API logs:

docker compose logs data-api

Follow logs in real-time:

docker compose logs -f

Configuration details

The Docker Compose setup includes the following configuration.

HCD service node configuration
Setting Value

Image

cp.icr.io/cp/cpd/ibm-datastax-hcd/hcd:1.2.4-ubi

Memory limit

3 GB

Max heap size

1536 MB

Cluster name

hcd-cluster

CQL port

9042

Platform

linux/amd64
Data API service configuration
Setting Value

Image

cp.icr.io/cp/cpd/ibm-datastax-hcd/data-api:v1.0.33

Memory limit

2 GB

HTTP port

8181

Contact points

hcd

This value must match the name of the HCD service in the compose.yaml file.

Local datacenter

dc1

Features

Document API, Table API, Vectorize

Customize the configuration

You can modify the compose.yaml and cassandra.yaml files to customize your installation.

Change memory limits

Edit the compose.yaml file to adjust memory limits:

services:
  hcd:
    mem_limit: 4G  # Increase HCD memory
    environment:
      - MAX_HEAP_SIZE=2048M  # Increase heap size

Change ports

Edit the compose.yaml file to use different ports:

services:
  hcd:
    ports:
      - 9043:9042  # Use port 9043 instead of 9042
  data-api:
    ports:
      - 8182:8181  # Use port 8182 instead of 8181

Modify settings

Edit the cassandra.yaml file to change database settings such as:

  • Replication settings

  • Compaction strategies

  • Cache sizes

  • Timeout values

For details on available settings, see cassandra.yaml configuration file.

Troubleshoot issues

Use these troubleshooting tips to resolve common issues.

Authentication fails

If you can’t authenticate with the ICR:

  • Verify your entitlement key is valid.

  • Ensure you’re using cp as the username.

  • Check that you have access to the HCD container images.

Contact IBM Support if you continue to experience issues.

Containers fail to start

Check the container logs:

docker compose logs

Common issues:

  • Insufficient memory: Increase Docker’s memory allocation to at least 6 GB total.

  • Port conflicts: Ensure ports 9042 and 8181 are available.

  • Disk space: Verify sufficient disk space is available (at least 10 GB free).

  • Image pull errors: Verify authentication with IBM Container Registry.

  • Platform mismatch: The images are built for linux/amd64. On Apple Silicon Macs, Docker/Podman uses emulation.

Can’t connect to CQL

Verify the HCD container is running and healthy:

docker compose ps
docker exec -it hcd-dev-hcd-1 nodetool status

Wait for the health check to pass. The initial startup can take 2-3 minutes.

Data API can’t connect to HCD

Ensure the HCD container is healthy before the Data API starts:

  1. Check the HCD container status:

    docker compose ps hcd

    You should see status "Up" and "healthy".

  2. If the HCD container is not healthy, check its logs:

    docker compose logs hcd

    Look for startup errors or initialization issues.

  3. Verify the HCD node is ready:

    docker exec -it hcd-dev-hcd-1 nodetool status

    You should see status "UN" (Up/Normal).

  4. If the Data API started before HCD was ready, restart it:

    docker compose restart data-api

    The Docker Compose configuration includes a health check dependency, but in some cases you might need to manually restart the Data API.

  5. Verify the Data API can reach HCD:

    docker compose logs data-api

    Look for successful connection messages to the contact point.

Performance issues on Apple Silicon

If you’re running on Apple Silicon (M1/M2/M3) Macs:

  • The linux/amd64 images run under emulation, which impacts performance.

  • Increase memory allocation in Docker Desktop settings.

Container names

The actual container names include the directory name as a prefix. If you created the installation in a directory called hcd-dev, the containers are named:

  • hcd-dev-hcd-1

  • hcd-dev-data-api-1

Adjust commands accordingly based on your directory name.

Next steps

Was this helpful?

Give Feedback

How can we improve the documentation?

© Copyright IBM Corporation 2026 | Privacy policy | Terms of use |  Manage Privacy Choices

Apache, Apache Cassandra, Cassandra, Apache Tomcat, Tomcat, Apache Lucene, Apache Solr, Apache Hadoop, Hadoop, Apache Pulsar, Pulsar, Apache Spark, Spark, Apache TinkerPop, TinkerPop, Apache Kafka and Kafka are either registered trademarks or trademarks of the Apache Software Foundation or its subsidiaries in Canada, the United States and/or other countries. Kubernetes is the registered trademark of the Linux Foundation.

General Inquiries: Contact IBM