Skip to content

Latest commit

 

History

History
395 lines (285 loc) · 38.9 KB

iceberg-rest-service.md

File metadata and controls

395 lines (285 loc) · 38.9 KB
Raw
title slug keywords license
Iceberg REST catalog service
/iceberg-rest-service
Iceberg REST catalog
This software is licensed under the Apache License version 2.

Background

The Apache Gravitino Iceberg REST Server follows the Apache Iceberg REST API specification and acts as an Iceberg REST catalog server.

Capabilities

  • Supports the Apache Iceberg REST API defined in Iceberg 1.5, and supports all namespace and table interfaces. The following interfaces are not implemented yet:
    • token
    • multi table transaction
    • pagination
  • Works as a catalog proxy, supporting Hive and JDBC as catalog backend.
  • Supports different storages like S3, HDFS, OSS, GCS and provides the capability to support other storages.
  • Supports OAuth2 and HTTPS.
  • Provides a pluggable metrics store interface to store and delete Iceberg metrics.

Server management

There are three deployment scenarios for Gravitino Iceberg REST server:

  • A standalone server in a standalone Gravitino Iceberg REST server package.
  • A standalone server in the Gravitino server package.
  • An auxiliary service embedded in the Gravitino server.

For detailed instructions on how to build and install the Gravitino server package, please refer to How to build and How to install. To build the Gravitino Iceberg REST server package, use the command ./gradlew compileIcebergRESTServer -x test. Alternatively, to create the corresponding compressed package in the distribution directory, use ./gradlew assembleIcebergRESTServer -x test. The Gravitino Iceberg REST server package includes the following files:

|── ...
└── distribution/gravitino-iceberg-rest-server
    |── bin/
    |   └── gravitino-iceberg-rest-server.sh    # Gravitino Iceberg REST server Launching scripts.
    |── conf/                                   # All configurations for Gravitino Iceberg REST server.
    |   ├── gravitino-iceberg-rest-server.conf  # Gravitino Iceberg REST server configuration.
    |   ├── gravitino-env.sh                    # Environment variables, etc., JAVA_HOME, GRAVITINO_HOME, and more.
    |   └── log4j2.properties                   # log4j configuration for the Gravitino Iceberg REST server.
    |   └── hdfs-site.xml & core-site.xml       # HDFS configuration files.
    |── libs/                                   # Gravitino Iceberg REST server dependencies libraries.
    |── logs/                                   # Gravitino Iceberg REST server logs. Automatically created after the server starts.

Apache Gravitino Iceberg REST catalog server configuration

There are distinct configuration files for standalone and auxiliary server: gravitino-iceberg-rest-server.conf is used for the standalone server, while gravitino.conf is for the auxiliary server. Although the configuration files differ, the configuration items remain the same.

Starting with version 0.6.0, the prefix gravitino.auxService.iceberg-rest. for auxiliary server configurations has been deprecated. If both gravitino.auxService.iceberg-rest.key and gravitino.iceberg-rest.key are present, the latter will take precedence. The configurations listed below use the gravitino.iceberg-rest. prefix.

Configuration to enable Iceberg REST service in Gravitino server.

Configuration item Description Default value Required Since Version
gravitino.auxService.names The auxiliary service name of the Gravitino Iceberg REST catalog service. Use iceberg-rest. (none) Yes 0.2.0
gravitino.iceberg-rest.classpath The classpath of the Gravitino Iceberg REST catalog service; includes the directory containing jars and configuration. It supports both absolute and relative paths, for example, iceberg-rest-server/libs, iceberg-rest-server/conf (none) Yes 0.2.0

Please note that, it only takes affect in gravitino.conf, you don't need to specify the above configurations if start as a standalone server.

HTTP server configuration

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.host The host of the Gravitino Iceberg REST catalog service. 0.0.0.0 No 0.2.0
gravitino.iceberg-rest.httpPort The port of the Gravitino Iceberg REST catalog service. 9001 No 0.2.0
gravitino.iceberg-rest.minThreads The minimum number of threads in the thread pool used by the Jetty web server. minThreads is 8 if the value is less than 8. Math.max(Math.min(Runtime.getRuntime().availableProcessors() * 2, 100), 8) No 0.2.0
gravitino.iceberg-rest.maxThreads The maximum number of threads in the thread pool used by the Jetty web server. maxThreads is 8 if the value is less than 8, and maxThreads must be greater than or equal to minThreads. Math.max(Runtime.getRuntime().availableProcessors() * 4, 400) No 0.2.0
gravitino.iceberg-rest.threadPoolWorkQueueSize The size of the queue in the thread pool used by Gravitino Iceberg REST catalog service. 100 No 0.2.0
gravitino.iceberg-rest.stopTimeout The amount of time in ms for the Gravitino Iceberg REST catalog service to stop gracefully. For more information, see org.eclipse.jetty.server.Server#setStopTimeout. 30000 No 0.2.0
gravitino.iceberg-rest.idleTimeout The timeout in ms of idle connections. 30000 No 0.2.0
gravitino.iceberg-rest.requestHeaderSize The maximum size of an HTTP request. 131072 No 0.2.0
gravitino.iceberg-rest.responseHeaderSize The maximum size of an HTTP response. 131072 No 0.2.0
gravitino.iceberg-rest.customFilters Comma-separated list of filter class names to apply to the APIs. (none) No 0.4.0

The filter in customFilters should be a standard javax servlet filter. You can also specify filter parameters by setting configuration entries in the style gravitino.iceberg-rest.<class name of filter>.param.<param name>=<value>.

Security

Gravitino Iceberg REST server supports OAuth2 and HTTPS, please refer to Security for more details.

Backend authentication

For JDBC backend, you can use the gravitino.iceberg-rest.jdbc.user and gravitino.iceberg-rest.jdbc.password to authenticate the JDBC connection. For Hive backend, you can use the gravitino.iceberg-rest.authentication.type to specify the authentication type, and use the gravitino.iceberg-rest.authentication.kerberos.principal and gravitino.iceberg-rest.authentication.kerberos.keytab-uri to authenticate the Kerberos connection. The detailed configuration items are as follows:

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.authentication.type The type of authentication for Iceberg rest catalog backend. This configuration only applicable for for Hive backend, and only supports Kerberos, simple currently. As for JDBC backend, only username/password authentication was supported now. simple No 0.6.0
gravitino.iceberg-rest.authentication.impersonation-enable Whether to enable impersonation for the Iceberg catalog false No 0.6.0
gravitino.iceberg-rest.authentication.kerberos.principal The principal of the Kerberos authentication (none) required if the value of gravitino.iceberg-rest.authentication.type is Kerberos. 0.6.0
gravitino.iceberg-rest.authentication.kerberos.keytab-uri The URI of The keytab for the Kerberos authentication. (none) required if the value of gravitino.iceberg-rest.authentication.type is Kerberos. 0.6.0
gravitino.iceberg-rest.authentication.kerberos.check-interval-sec The check interval of Kerberos credential for Iceberg catalog. 60 No 0.6.0
gravitino.iceberg-rest.authentication.kerberos.keytab-fetch-timeout-sec The fetch timeout of retrieving Kerberos keytab from authentication.kerberos.keytab-uri. 60 No 0.6.0

Storage

S3 configuration

Gravitino Iceberg REST service supports using static access-key-id and secret-access-key to access S3 data.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.io-impl The IO implementation for FileIO in Iceberg, use org.apache.iceberg.aws.s3.S3FileIO for S3. (none) No 0.6.0
gravitino.iceberg-rest.s3-access-key-id The static access key ID used to access S3 data. (none) No 0.6.0
gravitino.iceberg-rest.s3-secret-access-key The static secret access key used to access S3 data. (none) No 0.6.0
gravitino.iceberg-rest.s3-endpoint An alternative endpoint of the S3 service, This could be used for S3FileIO with any s3-compatible object storage service that has a different endpoint, or access a private S3 endpoint in a virtual private cloud. (none) No 0.6.0
gravitino.iceberg-rest.s3-region The region of the S3 service, like us-west-2. (none) No 0.6.0

For other Iceberg s3 properties not managed by Gravitino like s3.sse.type, you could config it directly by gravitino.iceberg-rest.s3.sse.type.

:::info To configure the JDBC catalog backend, set the gravitino.iceberg-rest.warehouse parameter to s3://{bucket_name}/${prefix_name}. For the Hive catalog backend, set gravitino.iceberg-rest.warehouse to s3a://{bucket_name}/${prefix_name}. Additionally, download the Iceberg AWS bundle and place it in the classpath of Iceberg REST server. :::

OSS configuration

Gravitino Iceberg REST service supports using static access-key-id and secret-access-key to access OSS data.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.io-impl The IO implementation for FileIO in Iceberg, use org.apache.iceberg.aliyun.oss.OSSFileIO for OSS. (none) No 0.6.0
gravitino.iceberg-rest.oss-access-key-id The static access key ID used to access OSS data. (none) No 0.7.0
gravitino.iceberg-rest.oss-secret-access-key The static secret access key used to access OSS data. (none) No 0.7.0
gravitino.iceberg-rest.oss-endpoint The endpoint of Aliyun OSS service. (none) No 0.7.0

For other Iceberg OSS properties not managed by Gravitino like client.security-token, you could config it directly by gravitino.iceberg-rest.client.security-token.

:::info Please set the gravitino.iceberg-rest.warehouse parameter to oss://{bucket_name}/${prefix_name}. Additionally, download the Aliyun OSS SDK and copy aliyun-sdk-oss-3.10.2.jar, hamcrest-core-1.1.jar, jdom2-2.0.6.jar in the classpath of Iceberg REST server, iceberg-rest-server/libs for the auxiliary server, libs for the standalone server. :::

GCS

Supports using google credential file to access GCS data.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.io-impl The io implementation for FileIO in Iceberg, use org.apache.iceberg.gcp.gcs.GCSFileIO for GCS. (none) No 0.6.0

For other Iceberg GCS properties not managed by Gravitino like gcs.project-id, you could config it directly by gravitino.iceberg-rest.gcs.project-id.

Please make sure the credential file is accessible by Gravitino, like using export GOOGLE_APPLICATION_CREDENTIALS=/xx/application_default_credentials.json before Gravitino Iceberg REST server is started.

:::info Please set gravitino.iceberg-rest.warehouse to gs://{bucket_name}/${prefix_name}, and download Iceberg gcp bundle and place it to the classpath of Gravitino Iceberg REST server, iceberg-rest-server/libs for the auxiliary server, libs for the standalone server. :::

HDFS configuration

You should place HDFS configuration file to the classpath of the Iceberg REST server, iceberg-rest-server/conf for Gravitino server package, conf for standalone Gravitino Iceberg REST server package. When writing to HDFS, the Gravitino Iceberg REST catalog service can only operate as the specified HDFS user and doesn't support proxying to other HDFS users. See How to access Apache Hadoop for more details.

:::info Builds with Hadoop 2.10.x. There may be compatibility issues when accessing Hadoop 3.x clusters. :::

Other storages

For other storages that are not managed by Gravitino directly, you can manage them through custom catalog properties.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.io-impl The IO implementation for FileIO in Iceberg, please use the full qualified classname. (none) No 0.6.0

To pass custom properties such as security-token to your custom FileIO, you can directly configure it by gravitino.iceberg-rest.security-token. security-token will be included in the properties when the initialize method of FileIO is invoked.

:::info Please set the gravitino.iceberg-rest.warehouse parameter to {storage_prefix}://{bucket_name}/${prefix_name}. Additionally, download corresponding jars in the classpath of Iceberg REST server, iceberg-rest-server/libs for the auxiliary server, libs for the standalone server. :::

Catalog backend configuration

:::info The Gravitino Iceberg REST catalog service uses the memory catalog backend by default. You can specify a Hive or JDBC catalog backend for production environment. :::

Hive backend configuration

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.catalog-backend The Catalog backend of the Gravitino Iceberg REST catalog service. Use the value hive for a Hive catalog. memory Yes 0.2.0
gravitino.iceberg-rest.uri The Hive metadata address, such as thrift://127.0.0.1:9083. (none) Yes 0.2.0
gravitino.iceberg-rest.warehouse The warehouse directory of the Hive catalog, such as /user/hive/warehouse-hive/. (none) Yes 0.2.0
gravitino.iceberg-rest.catalog-backend-name The catalog backend name passed to underlying Iceberg catalog backend. Catalog name in JDBC backend is used to isolate namespace and tables. hive for Hive backend, jdbc for JDBC backend, memory for memory backend No 0.5.2

JDBC backend configuration

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.catalog-backend The Catalog backend of the Gravitino Iceberg REST catalog service. Use the value jdbc for a JDBC catalog. memory Yes 0.2.0
gravitino.iceberg-rest.uri The JDBC connection address, such as jdbc:postgresql://127.0.0.1:5432 for Postgres, or jdbc:mysql://127.0.0.1:3306/ for mysql. (none) Yes 0.2.0
gravitino.iceberg-rest.warehouse The warehouse directory of JDBC catalog. Set the HDFS prefix if using HDFS, such as hdfs://127.0.0.1:9000/user/hive/warehouse-jdbc (none) Yes 0.2.0
gravitino.iceberg-rest.catalog-backend-name The catalog name passed to underlying Iceberg catalog backend. Catalog name in JDBC backend is used to isolate namespace and tables. jdbc for JDBC backend No 0.5.2
gravitino.iceberg-rest.jdbc.user The username of the JDBC connection. (none) Yes 0.2.0
gravitino.iceberg-rest.jdbc.password The password of the JDBC connection. (none) Yes 0.2.0
gravitino.iceberg-rest.jdbc-initialize Whether to initialize the meta tables when creating the JDBC catalog. true No 0.2.0
gravitino.iceberg-rest.jdbc-driver com.mysql.jdbc.Driver or com.mysql.cj.jdbc.Driver for MySQL, org.postgresql.Driver for PostgreSQL. (none) Yes 0.3.0

If you have a JDBC Iceberg catalog prior, you must set catalog-backend-name to keep consistent with your Jdbc Iceberg catalog name to operate the prior namespace and tables.

:::caution You must download the corresponding JDBC driver to the iceberg-rest-server/libs directory. :::

Custom backend configuration

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.catalog-backend The Catalog backend of the Gravitino Iceberg REST catalog service. Use the value custom for a Custom catalog. memory Yes 0.2.0
gravitino.iceberg-rest.catalog-backend-impl The fully-qualified class name of a custom catalog implementation, only worked if catalog-backend is custom. (none) No 0.7.0

If you want to use a custom Iceberg Catalog as catalog-backend, you can add a corresponding jar file to the classpath and load a custom Iceberg Catalog implementation by specifying the catalog-backend-impl property.

View support

You could access the view interface if using JDBC backend and enable jdbc.schema-version property.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.jdbc.schema-version The schema version of JDBC catalog backend, setting to V1 if supporting view operations. (none) NO 0.7.0

Multi catalog support

The Gravitino Iceberg REST server supports multiple catalogs and offers a configuration-based catalog management system.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.catalog-provider Catalog provider class name, you can develop a class that implements IcebergTableOpsProvider and add the corresponding jar file to the Iceberg REST service classpath directory. config-based-provider No 0.7.0
Configuration based catalog provider

When using a configuration based catalog provider, you can configure the default catalog with gravitino.iceberg-rest.catalog.<param name>=<value>. For specific catalogs, use the format gravitino.iceberg-rest.catalog.<catalog name>.<param name>=<value>.

For instance, you can configure three different catalogs, the default catalog and the specific hive_backend and jdbc_backend catalogs separately.

gravitino.iceberg-rest.catalog-backend = jdbc
gravitino.iceberg-rest.uri = jdbc:postgresql://127.0.0.1:5432
gravitino.iceberg-rest.warehouse = hdfs://127.0.0.1:9000/user/hive/warehouse-postgresql
...
gravitino.iceberg-rest.catalog.hive_backend.catalog-backend = hive
gravitino.iceberg-rest.catalog.hive_backend.uri = thrift://127.0.0.1:9084
gravitino.iceberg-rest.catalog.hive_backend.warehouse = /user/hive/warehouse-hive/
...
gravitino.iceberg-rest.catalog.jdbc_backend.catalog-backend = jdbc
gravitino.iceberg-rest.catalog.jdbc_backend.uri = jdbc:mysql://127.0.0.1:3306/
gravitino.iceberg-rest.catalog.jdbc_backend.warehouse = hdfs://127.0.0.1:9000/user/hive/warehouse-mysql
...

You can access different catalogs by setting the prefix to the specific catalog name in the Iceberg REST client configuration. The default catalog will be used if you do not specify a prefix. For instance, consider the case of SparkSQL.

./bin/spark-sql -v \
...
--conf spark.sql.catalog.default_rest_catalog.type=rest  \
--conf spark.sql.catalog.default_rest_catalog.uri=http://127.0.0.1:9001/iceberg/ \
...
--conf spark.sql.catalog.hive_backend_catalog.type=rest  \
--conf spark.sql.catalog.hive_backend_catalog.uri=http://127.0.0.1:9001/iceberg/ \
--conf spark.sql.catalog.hive_backend_catalog.prefix=hive_backend \
...
--conf spark.sql.catalog.jdbc_backend_catalog.type=rest  \
--conf spark.sql.catalog.jdbc_backend_catalog.uri=http://127.0.0.1:9001/iceberg/ \
--conf spark.sql.catalog.jdbc_backend_catalog.prefix=jdbc_backend \
...
Gravitino server based catalog provider

When using a Gravitino server based catalog provider, you can leverage Gravitino to support dynamic catalog management for the Iceberg REST server.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.gravitino-uri The uri of Gravitino server address, only worked if catalog-provider is gravitino-based-provider. (none) No 0.7.0
gravitino.iceberg-rest.gravitino-metalake The metalake name that gravitino-based-provider used to request to Gravitino, only worked if catalog-provider is gravitino-based-provider. (none) No 0.7.0
gravitino.iceberg-rest.catalog-cache-eviction-interval-ms Catalog cache eviction interval. 3600000 No 0.7.0 
gravitino.iceberg-rest.catalog-cache-eviction-interval-ms = 300000
gravitino.iceberg-rest.catalog-provider = gravitino-based-provider
gravitino.iceberg-rest.gravitino-uri = http://127.0.0.1:8090
gravitino.iceberg-rest.gravitino-metalake = test

Other Apache Iceberg catalog properties

You can add other properties defined in Iceberg catalog properties. The clients property for example:

Configuration item Description Default value Required
gravitino.iceberg-rest.clients The client pool size of the catalog. 2 No

:::info catalog-impl has no effect. :::

Apache Iceberg metrics store configuration

Gravitino provides a pluggable metrics store interface to store and delete Iceberg metrics. You can develop a class that implements org.apache.gravitino.iceberg.service.metrics.IcebergMetricsStore and add the corresponding jar file to the Iceberg REST service classpath directory.

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.metricsStore The Iceberg metrics storage class name. (none) No 0.4.0
gravitino.iceberg-rest.metricsStoreRetainDays The days to retain Iceberg metrics in store, the value not greater than 0 means retain forever. -1 No 0.4.0
gravitino.iceberg-rest.metricsQueueCapacity The size of queue to store metrics temporally before storing to the persistent storage. Metrics will be dropped when queue is full. 1000 No 0.4.0

Misc configurations

Configuration item Description Default value Required Since Version
gravitino.iceberg-rest.extension-packages Comma-separated list of Iceberg REST API packages to expand. (none) No 0.7.0

Starting the Iceberg REST server

To start as an auxiliary service with Gravitino server:

./bin/gravitino.sh start

To start a standalone Gravitino Iceberg REST catalog server:

./bin/gravitino-iceberg-rest-server.sh start

To verify whether the service has started:

curl  http://127.0.0.1:9001/iceberg/v1/config

Normally you will see the output like {"defaults":{},"overrides":{}}%.

Exploring the Apache Gravitino Iceberg REST catalog service with Apache Spark

Deploying Apache Spark with Apache Iceberg support

Follow the Spark Iceberg start guide to set up Apache Spark's and Apache Iceberg's environment.

Starting the Apache Spark client with the Apache Iceberg REST catalog

Configuration item Description
spark.sql.catalog.${catalog-name}.type The Spark catalog type; should set to rest.
spark.sql.catalog.${catalog-name}.uri Spark Iceberg REST catalog URI, such as http://127.0.0.1:9001/iceberg/.

For example, we can configure Spark catalog options to use Gravitino Iceberg REST catalog with the catalog name rest.

./bin/spark-sql -v \
--packages org.apache.iceberg:iceberg-spark-runtime-3.4_2.12:1.3.1 \
--conf spark.sql.extensions=org.apache.iceberg.spark.extensions.IcebergSparkSessionExtensions \
--conf spark.sql.catalog.rest=org.apache.iceberg.spark.SparkCatalog  \
--conf spark.sql.catalog.rest.type=rest  \
--conf spark.sql.catalog.rest.uri=http://127.0.0.1:9001/iceberg/

You may need to adjust the Iceberg Spark runtime jar file name according to the real version number in your environment. If you want to access the data stored in cloud, you need to download corresponding jars (please refer to the cloud storage part) and place it in the classpath of Spark, no extra config is needed because related properties is transferred from Iceberg REST server to Iceberg REST client automatically.

For other storages not managed by Gravitino, the properties wouldn't transfer from the server to client automatically, if you want to pass custom properties to initialize FileIO, you could add it by spark.sql.catalog.${iceberg_catalog_name}.${configuration_key} = {property_value}.

Exploring Apache Iceberg with Apache Spark SQL

// First change to use the `rest` catalog
USE rest;
CREATE DATABASE IF NOT EXISTS dml;
CREATE TABLE dml.test (id bigint COMMENT 'unique id') using iceberg;
DESCRIBE TABLE EXTENDED dml.test;
INSERT INTO dml.test VALUES (1), (2);
SELECT * FROM dml.test;

Docker instructions

You could run Gravitino Iceberg REST server though docker container:

docker run -d -p 9001:9001 apache/gravitino-iceberg-rest:0.6.0

Or build it manually to add custom logics:

sh ./dev/docker/build-docker.sh --platform linux/arm64 --type iceberg-rest-server --image apache/gravitino-iceberg-rest --tag 0.6.0

You could try Spark with Gravitino REST catalog service in our playground.