Trino Connector Integration Developer Guide

This document targets developers who need to integrate Trino Connector plugins with Apache Doris. It covers how to compile a Connector plugin, configure FE / BE, and create and use a trino-connector Catalog. The Kafka Connector is used as the main example, with additional configuration samples for common plugins such as Hive, MySQL, and BigQuery.

Contents

• Background: when to use the Trino Connector framework
• Step 1 - Compile the Connector plugin: Kafka as an example
• Step 2 - Configure FE / BE: plugin directory and lazy-loading behavior
• Step 3 - Create and use the Catalog: CREATE CATALOG and property mapping
• Common Connector configuration examples: Hive / MySQL / Kafka / BigQuery
• FAQ and common errors

---

Background

Starting from version 3.0, Doris supports integration with Trino Connector plugins. With the rich ecosystem of Trino Connector plugins and the trino-connector Catalog feature in Doris, Doris can connect to more data sources.

Scenario	Recommended approach
Mainstream data sources such as Hive, Iceberg, Hudi, Paimon, JDBC	Use the Doris built-in Catalog first, for better performance, stability, and compatibility
Other data sources that Doris does not yet integrate	Use the Trino Connector compatibility framework to quickly connect to more data sources

Note: Trino is open-source software licensed under the Apache License 2.0, provided by the Trino Software Foundation. For more information, visit the Trino official site.

---

Step 1 - Compile the Connector plugin

Trino does not provide officially compiled Connector plugins, so you need to compile the required Connector plugins yourself.

Note: Doris currently uses the 435 version of the trino-main package, so it is best to compile the 435 version of the Connector plugin. Connector plugins of other versions may have compatibility issues. If you encounter problems, please report them to the Apache Doris community.

Take the Kafka plugin as an example:

# 1. Clone the Trino source code
git clone https://github.com/trinodb/trino.git

# 2. Check out version 435
git checkout 435

# 3. Enter the Kafka plugin source directory
cd trino/plugin/trino-kafka

# 4. Compile the Kafka plugin
mvn clean install -DskipTests

After compilation, a target/trino-kafka-435 directory is generated under trino/plugin/trino-kafka/.

Note: Each Connector plugin is a subdirectory, not a single jar file.

---

Step 2 - Configure FE / BE

After preparing the Kafka Connector plugin, you need to configure the plugin directory in fe.conf and be.conf so that Doris can locate the plugin.

Assuming the trino-kafka-435 directory prepared in the previous step is placed under /path/to/connectors, configure as follows:

File	Configuration item	Default value
fe.conf	trino_connector_plugin_dir=/path/to/connectors	${Doris_HOME}/fe/connectors
be.conf	trino_connector_plugin_dir=/path/to/connectors	${Doris_HOME}/be/connectors

Lazy-loading behavior: Doris loads Trino Connector plugins lazily. The first time you use the trino-connector Catalog feature, you do not need to restart FE / BE nodes; Doris loads the plugin automatically. However, plugins are loaded only once, so after the plugin directory contents change, you must restart FE / BE nodes for the changes to take effect.

---

Step 3 - Create and use the Catalog

3.1 Create a Trino-Connector Catalog

create catalog kafka_tpch properties (
    "type"="trino-connector",
    -- The following four properties come from Trino and match the properties in Trino's etc/catalog/kafka.properties, but each must be prefixed with "trino."
    "trino.connector.name"="kafka",
    "trino.kafka.table-names"="tpch.customer,tpch.orders,tpch.lineitem,tpch.part,tpch.partsupp,tpch.supplier,tpch.nation,tpch.region",
    "trino.kafka.nodes"="localhost:9092",
    "trino.kafka.table-description-dir" = "/mnt/datadisk1/fangtiewei"
);

Property descriptions:

Property	Description
type	Catalog type. Must be set to trino-connector
trino.connector.name	Name of the Trino Connector, for example kafka
trino.kafka.table-names and others	These properties come from Trino, all prefixed with trino.. See Kafka connector

Different Connector plugins require different properties. See the Trino official documentation: Connectors.

3.2 Use the Catalog

Once the Trino-Connector Catalog is created, using it is no different from using any other Catalog. Switch to the Catalog with switch kafka_tpch and query data from the corresponding data source:

switch kafka_tpch;
show databases;
select * from tpch.customer limit 10;

---

Common Connector configuration examples

Hive

create catalog emr_hive properties (
    "type"="trino-connector",
    "trino.connector.name"="hive",
    "trino.hive.metastore.uri"="thrift://ip:port",
    "trino.hive.config.resources"="/path/to/core-site.xml,/path/to/hdfs-site.xml"
);

Notes for the Hive plugin:

• You must add the Hadoop user to the JVM parameters: -DHADOOP_USER_NAME=user. This can be appended to the JAVA_OPTS_FOR_JDK_17 parameter in fe.conf / be.conf, for example JAVA_OPTS_FOR_JDK_17="...-DHADOOP_USER_NAME=user".

MySQL

create catalog trino_mysql properties (
    "type"="trino-connector",
    "trino.connector.name"="mysql",
    "trino.connection-url" = "jdbc:mysql://ip:port",
    "trino.connection-user" = "user",
    "trino.connection-password" = "password"
);

Notes for the MySQL plugin:

• If you see the error Unknown or incorrect time zone: 'Asia/Shanghai', append -Duser.timezone=Etc/GMT-8 to the JVM startup parameters. You can add it at the end of the JAVA_OPTS_FOR_JDK_17 parameter in fe.conf / be.conf.

Kafka

create catalog kafka properties (
    "type"="trino-connector",
    "trino.connector.name"="kafka",
    "trino.kafka.nodes"="localhost:9092",
    "trino.kafka.table-description-supplier"="CONFLUENT",
    "trino.kafka.confluent-schema-registry-url"="http://localhost:8081",
    "trino.kafka.hide-internal-columns" = "false"
);

BigQuery

create catalog bigquery_catalog properties (
    "type"="trino-connector",
    "trino.connector.name"="bigquery",
    "trino.bigquery.project-id"="steam-circlet-388406",
    "trino.bigquery.credentials-file"="/path/to/application_default_credentials.json"
);

---

FAQ and common errors

Q: I modified a plugin under /path/to/connectors, why does Doris not pick up the change?

A: Doris uses lazy loading, but plugins are loaded only once. After modifying a plugin, you must restart FE / BE nodes.

Q: Startup reports a plugin version incompatibility error?

A: Confirm that you checked out the Trino 435 branch when compiling. Doris currently uses version 435 of trino-main, and other versions may have compatibility issues.

Q: Hive plugin queries report permission errors or fail to access HDFS?

A: Check whether JAVA_OPTS_FOR_JDK_17 has -DHADOOP_USER_NAME=user set, and whether the core-site.xml / hdfs-site.xml files referenced by trino.hive.config.resources are readable.

Q: MySQL plugin reports Unknown or incorrect time zone: 'Asia/Shanghai'?

A: Append -Duser.timezone=Etc/GMT-8 to JAVA_OPTS_FOR_JDK_17.

Q: Should I use Trino Connector for Hive, Iceberg, Hudi, and similar data sources?

A: Not recommended. The Doris built-in Catalog offers better performance, stability, and compatibility for these data sources. The Trino Connector framework is positioned to fill gaps for data sources that Doris does not yet integrate.