Launch an Amazon EMR cluster with Trino

Choose AWS Glue in the console and create a table, based on your source data in Amazon S3. A table in the AWS Glue Data Catalog is the metadata definition for the data. It makes sense in this context to create the table manually, creating columns as you like, from your source data. For more information about creating tables in AWS Glue from semi-structured data in Amazon S3, see Creating tables using the console in the AWS Glue User Guide.

Set your configuration as part of cluster creation. Select the Configurations tab. Configurations are optional specifications for your cluster. When you enter a configuration, add JSON like the following sample, which instructs Trino to use the AWS Glue Data Catalog as its external Hive metastore for table metadata:

{
    "classification": "trino-connector-hive",
    "properties": {
        "hive.metastore": "glue"
    }
}

Alternatively, you can apply configurations in the Software settings section when you create a cluster.

Additionally, you can set up other connector types, such as for connecting with Apache Iceberg. For more information, see Use an Iceberg cluster with Trino in the Amazon EMR Release Guide. Configuring additional settings is optional.

In the AWS console, select Amazon EMR from the services. When you choose Amazon EMR, if you have existing clusters, your EMR on EC2 clusters are listed.

Choose Create cluster. From here, you start the process for building a cluster.

Give your cluster a name and choose an Amazon EMR release. You can choose the most current release for the tutorial.

Choose the Trino bundle, which has the Trino application pre-selected. Bundles are set up for convenience when you know the purpose for the cluster ahead of time. Otherwise, you can simply select the check box for Trino.

For Cluster configuration, choose Uniform instance groups. Go ahead and remove additional instance groups.

Choose an Instance type. Generally we recommend you choose an instance type with at least 16 GiB memory. Also, for Cluster scaling and provisioning choose Set cluster size manually.

At this point, set your Hive metastore configuration to point to AWS Glue. This is detailed in the section Using a Hive connector to make data available for querying. Complete this before you build the cluster.

Choose Create cluster. It can take a few minutes to finish.

The steps here don't cover all of the configuration steps in detail. More information about setting up a cluster is available at Plan, configure and launch Amazon EMR clusters.