The Metadata Access Service is an application service that allows the Developer tool to access the Hadoop environment to import and preview metadata. If the domain uses non-Kerberos authentication, you can create and configure the Metadata Access Service. If the domain uses Kerberos authentication, don't choose to create the Metadata Access Service.
Use the Administrator tool or the
infacmd
command line program to create and administer the Metadata Access Service. You can create one or more Metadata Access Services on a node. You can use one Metadata Access Service to import and preview metadata from multiple Hadoop distributions. The Metadata Access Service creates a new process to connect to each Hadoop distribution. Based on your license, the Metadata Access Service can be highly available.
If your domain has only one Metadata Access Service, the Developer tool uses the same Metadata Access Service by default to fetch metadata from multiple Hadoop distributions. If your domain has more than one Metadata Access Service, you must select a default Metadata Access Service for the Developer tool to connect to the default Metadata Access Service. Dynamic mappings do not use the Metadata Access Service.
The Developer tool does not use the Metadata Access Service to access the Databricks environment.
HBase, HDFS, Hive, and MapR-DB connections use the Metadata Access Service when you import an object from a Hadoop cluster. Google Cloud Storage connection uses Metadata Access Service to import metadata from files in Google Cloud Storage. Create and configure a Metadata Access Service before you create Google Cloud Storage, HBase, HDFS, Hive, and MapR-DB connections.