OpenMetadata Server | Developer Guide & Setup

Build and run metadata server locally with this guide, including environment setup and server configuration tips.

OpenMetadata Server

Learn how to run the OpenMetadata server in development mode by using Docker and IntelliJ.

Prerequisites

• For an easy installation of MySQL and ES, just install Docker on your local machine and run the following commands from the top-level directory

docker compose -f docker/development/docker-compose.yml up mysql elasticsearch --build -d

• For an easy installation of PostgreSQL and ES, just install Docker on your local machine and run the following commands from the top-level directory

docker compose -f docker/development/docker-compose-postgres.yml up postgresql opensearch --build -d

Building

The following commands must be run from the top-level directory.

mvn clean install

This will run all tests, which can take up to an hour. Use -DskipTests to skip unit tests if you’re only interested in running the server.

Create a distribution (packaging)

You can create a distribution as follows.

$ mvn clean install

The binaries will be created at:

openmetadata-dist/target/open-metadata-<version>.pom
openmetadata-dist/target/open-metadata-<version>.tar.gz

Bootstrap Database

Extract the distribution tar.gz file created on the previous step and run the following command

cd open-metadata-<version>
sh bootstrap/openmetadata-ops.sh drop-create

Why Needed: The script simplifies the setup and execution of OpenMetadata by handling Java dependencies, configurations, and environment variables automatically.

Note To understand more about openmetadata-ops.sh file click here

Running the OpenMetadata server

You can run the OpenMetadata server directly

cd open-metadata-<version>
sh bin/openmetadata-server-start.sh conf/openmetadata.yaml

Using IntelliJ IDEA

Add a new Run/Debug configuration like the below screenshot.

1. Click on Intellij - Run menu
2. Click on “Edit Configurations”
3. Click + sign and Select Application and make sure your config looks similar to the below image

Note that if you want to test the connectors (Ingestion) while having the server running locally and the ingestion running as a container, you will need to ensure that both components can properly talk to each other. To do so, pass the following environment variable to IntelliJ when starting the server:

SERVER_HOST_API_URL=http://host.docker.internal:8585/api

Add missing dependency

Right-click on openmetadata-service

Click on “Open Module Settings”

Go to “Dependencies”

Click “+” at the bottom of the dialog box and click “Add”

Click on Library

In that list look for “jersey-client:2.25.1”

Select it and click “OK”.

We also need to set the folder ‘generated-resources’ in some module’s target folder as “source” folder. IntelliJ IDEA mark target folder as “excluded” by default, we could change it in the module setting. The openmetadata-spec and openmetadata-java-client modules have generated code, need to be changed.

Now run/debug the application.

Load Sample Data into the database

With the OpenMetadata service up and running, run the following commands from the top-level directory

python3 -m venv venv
source venv/bin/activate
make install_dev generate
cd ingestion
pip install -e '.[sample-data, elasticsearch]'
metadata ingest -c ./pipelines/sample_data.json
metadata usage -c ./pipelines/sample_usage.json

• You are now ready to explore the app by going to http://localhost:8585 *If the web page doesn’t work as intended, please take a look at the troubleshooting steps here

Troubleshooting

• If you see blank page at http://localhost:8585, please check the logs at logs/openmetadata.log. You might be encountering one of the following errors:
  • connection refused or unreachable - please confirm that MySQL and ES are reachable outside of docker by running docker ps and checking that ports 3306 and 9200 are listening on 0.0.0.0
    • If ElasticSearch in Docker on Mac is crashing, try changing Preferences -> Resources -> Memory to 4GB
    • If ElasticSearch logs show high disk watermark [90%] exceeded, try changing Preferences -> Resources -> Disk Image Size to at least 16GB
  • Public Key Retrieval is not allowed - verify that the JDBC connect URL in conf/openmetadata.yaml is configured with the parameter allowPublicKeyRetrieval=true
  • Browser console shows javascript errors, try doing a clean build. Some npm packages may not have been built properly.

Coding Style

1. Configure IntelliJ to disable the [wild-card imports] (https://www.jetbrains.com/help/idea/creating-and-optimizing-imports.html#disable-wildcard-imports)