Integration Test

    It is used to define the SQL to be tested and the assertion data of the test results.

    Each case defines one SQL, which can define multiple database execution types.

    Test environment

    It is used to set up the database and ShardingSphere-Proxy environment for running test cases. The environment is classified into environment preparation mode, database type, and scenario.

    Environment preparation mode is divided into Native and Docker, and Embed type will be supported in the future.

    • Native environment is used for test cases to run directly in the test environment provided by the developer, suitable for debugging scenarios;
    • Docker environment is directly built when Maven runs the Docker-Compose plug-in. It is suitable for cloud compilation environment and testing ShardingSphere-Proxy, such as GitHub Action;
    • Embed environment is built when the test framework automatically builds embedded MySQL. It is suitable for the local environment test of ShardingSphere-JDBC.

    Currently, the Native environment is adopted by default, and ShardingSphere-JDBC + H2 database is used to run test cases. Maven’s parameter specifies how the Docker environment is run. In the future, ShardingSphere-JDBC + MySQL of the Embed environment will be adopted to replace the default environment type used when Native executes test cases.

    Database types currently support MySQL, PostgreSQL, SQLServer, and Oracle, and test cases can be executed using ShardingSphere-JDBC or ShardingSphere-Proxy.

    Scenarios are used to test the supporting rules of ShardingSphere. Currently, data sharding and read/write splitting and other related scenarios are supported, and the combination of scenarios will be improved continuously in the future.

    It is used to read test cases in batches and execute and assert test results line by line.

    The test engine arranges test cases and environments to test as many scenarios as possible with the fewest test cases.

    Each SQL generates a test report in the combination of database type * access port type * SQL execution mode * JDBC execution mode * Scenario. Currently, each dimension is supported as follows:

    • Database types: H2, MySQL, PostgreSQL, SQLServer, and Oracle;

    • SQL execution modes: Statement and PreparedStatement;

    • JDBC execution modes: execute and executeQuery/executeUpdate;

    • Scenarios: database shards, table shards, read/write splitting and sharding + read/write splitting

    Therefore, one SQL will drive Database type (5) * Access port type (2) * SQL execution mode (2) * JDBC execution mode (2) * Scenario (4) = 160 test cases to be run to achieve the pursuit of high quality.

    User Guide

    Module path:test/e2e/suite

    Test case configuration

    SQL test case is in resources/cases/${SQL-TYPE}/${SQL-TYPE}-integration-test-cases.xml.

    The case file format is as follows:

    The lookup rule of expected-data-fileis as follows:

    1. Find the file dataset\${SCENARIO_NAME}\${DATABASE_TYPE}\${dataset_file}.xml in the same level directory;
    2. Find the file dataset\${SCENARIO_NAME}\${dataset_file}.xml in the same level directory;
    3. Find the file dataset\${dataset_file}.xml in the same level directory;
    4. Report an error if none of them are found.

    The assertion file format is as follows:

    ${SCENARIO-TYPE} Refers to the scenario name used to identify a unique scenario during the test engine run. ${DATABASE-TYPE} refers to the database types.

    Native environment configuration

    Directory: src/test/resources/env/${SCENARIO-TYPE}

    • : data source configuration;
    • rules.yaml: rule configuration;
    • dataset.xml: initialize the data;
    • init-sql\${DATABASE-TYPE}\init.sql: initialize the database and table structure;
    • authority.xml: to be supplemented.

    Docker environment configuration

    • docker-compose.yml: Docker-Compose config files, used for Docker environment startup;
    • proxy/conf/config-${SCENARIO-TYPE}.yaml: rule configuration。

    **The Docker environment configuration provides a remote debugging port for ShardingSphere-Proxy. You can find the second exposed port for remote debugging in shardingsphere-proxy of the docker-comemage. yml file. **

    Run the test engine

    Configure the running environment of the test engine

    Control the test engine by configuring src/test/resources/env/engine-env.properties.

    All attribute values can be dynamically injected via Maven command line .

    Run debugging mode

    • Standard test engine Run org.apache.shardingsphere.test.integration.engine.${SQL-TYPE}.General${SQL-TYPE}E2EIT to start the test engines of different SQL types.

    • Batch test engine Run org.apache.shardingsphere.test.integration.engine.dml.BatchDMLE2EIT to start the batch test engine for the test addBatch() provided for DML statements.

    • Additional test engine Run org.apache.shardingsphere.test.integration.engine.${SQL-TYPE}.Additional${SQL-TYPE}E2EIT to start the test engine with more JDBC method calls. Additional test engines need to be enabled by setting it.run.additional.cases=true.

    Run Docker mode

    Run the above command to build a Docker mirror apache/shardingsphere-proxy-test:latest used for integration testing. If you only modify the test code, you can reuse the existing test mirror without rebuilding it. Skip the mirror building and run the integration testing directly with the following command:

    Remote debug Proxy code in Docker container

    Remote debug Proxy started by docker image

    E2E Test Proxy image opens the 3308 port by default for remote debugging of the instance in the container. Use the following method to connect and debug the Proxy code in the container with IDE tools such as IDEA:

    IDEA -> Run -> Edit Configurations -> Add New Configuration -> Remote JVM Debug

    Edit the corresponding information:

    • Name: A descriptive name, such as e2e-debug.
    • Host: A IP that can access docker, such as 127.0.0.1
    • Port: debugging port 3308.
    • use module classpath: The root directory of the project shardingsphere.

    After editing the above information, run Run -> Run -> e2e-debug in IDEA to start the remote debug of IDEA.

    Remote debug Proxy started by Testcontainer
    • Set a breakpoint in the relevant startup class of Testcontainer, for example, after the line containerComposer.start(); in BaseE2EIT#setUp() in the suite test, at this time, the relevant containers must have been started.
    • Access breakpoint debugging mode through shortcut key Alt + F8, and view the port mapped by the 3308 mapping of the Proxy object under the containerComposer (the external mapping port of Testcontainer is random). For example, the expression ((ShardingSphereProxyClusterContainer)((java.util.LinkedList)((ITContainers)((ClusterContainerComposer)containerComposer).containers).dockerContainers).getLast()).getMappedPort(3308) get the mapped random port 51837.(or get mapped port by docker ps)
    • See the Remote debug Proxy started by docker image method, set the Name, Host, Port, and use the port got in previous step.

    After editing the above information, run Run -> Run -> e2e-debug in IDEA to start the remote debug of IDEA.

    Notice

    1. In order to ensure the integrity and legibility of the test data, 10 database shards and 10 table shards are used in the sharding of the integration testing, which takes a long time to run the test cases completely.