Upgrade from 0.13 to 1.0

Upgrading from 0.13 to 1.0 can be done directly on your instances. This page will list few general details you should take into consideration when running the upgrade.

  • Postgres:

    • The database field for Postgres connector has been made a required field as initial database is required to connect to Postgres.
    • Ingest All Database field has been added since earlier when database field was empty we were ingesting all the databases but since now we have made the database field required you need to enable Ingest All Database flag in case you want to ingest all the available database in OpenMetadata.
  • BigQuery:

    • Removed: partitionQueryDuration, partitionQuery & partitionField
      • since these fields are not used during the ingestion and has already being moved to profiler hence we have removed these fields.
  • Salesforce:

    • Removed: scheme, hostPort, supportsProfiler & supportsQueryComment
      • These fields are related to a SQLAlchemy based connector and since Salesforce is based on APIs and not SQLAlchemy, these fields has been removed
  • Superset:

    • Removed: connectionOptions
      • This field was not being used anywhere, hence removed.

With Release 1.0.0, JWT Authentication endpoint needs to be updated from {OPENMETADATA_HOST_NAME}/api/v1/config/jwks to {OPENMETADATA_HOST_NAME}/api/v1/system/config/jwks. This is required as part of API Endpoints. The Environment variable name is AUTHENTICATION_PUBLIC_KEYS. It expects list of URLs. One of the URL will be for OpenMetadata JWT Authentication Endpoint.

The new section on the openmetadata.yaml configuration for the Pipeline Service Client has been updated.

We now have a pipelineServiceClientConfiguration, instead of the old airflowConfiguration.

Most existing environment variables remain the same, except for these three:


When upgrading, make sure to update the environment variables and, if working on Bare Metal, make sure to use the updated openmetadata.yaml.

  • When configuring Bots, JWT tokens will be the preferred method of authentication. Any existing SSO-based service accounts will continue to work on 1.0, but will be fully deprecated on future releases.
  • As we added the new Impala connector, We will remove the impala scheme from Hive in the next release.

The following endpoints have been renamed in 1.0

Previous EndpointNew Endpoint

The SampleData service has been deprecated. It is now a CustomConnector. If you have some entities in SampleData, please DELETE the service if you don’t want to keep them, or we can help you migrate them to a Custom Connector.

Note that this service type was mostly used on quickstarts and tests to add some example assets into OpenMetadata. This should be transparent for most of the users.

We are deprecating the Location Entity in favor of the Containers and new Storage Service:

  • Dropping the location_entity table,
  • Removing the Location APIs.

If you did not have any custom implementation, this was partially used in the Glue Database connector. However, the information was not being actively shown.

If you had custom implementations on top of the Location APIs, reach out to us, and we can help migrate to the new Storage Services.

The endpointURL property is now formatted as a proper URI, e.g., http://something.com. If you have added this configuration in your connectors, please update the endpointURL format with the right scheme.

Note that this property is OPTIONAL, and for the most cases it will either be left blank or already configured with the right format for it to work properly, e.g., s3://....

  • metadata.test_suite.*: this submodule has been renamed metadata.data_quality.*. You can view the full change here
  • metadata.orm_profiler.*: this submodule has been renamed metadata.profiler.*. You can view the full change here