[ https://issues.apache.org/jira/browse/FLINK-30790?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Martijn Visser reassigned FLINK-30790: -------------------------------------- Assignee: João Boto > [Connectors/Jdbc] Refactor of testings > -------------------------------------- > > Key: FLINK-30790 > URL: https://issues.apache.org/jira/browse/FLINK-30790 > Project: Flink > Issue Type: Improvement > Components: Connectors / JDBC > Reporter: João Boto > Assignee: João Boto > Priority: Major > Labels: pull-request-available > > This is one attempt to get better results executing tests on jdbc connector.. > In current master branch it takes like 15m to execute all tests (it will vary > from 10m to 20m) > > {code:java} > NFO] Reactor Summary for Flink : Connectors : JDBC : Parent 3.1-SNAPSHOT: > 40943[INFO] > 40944[INFO] Flink : Connectors : JDBC : Parent ................. SUCCESS [ > 22.019 s] > 40945[INFO] Flink : Connectors : JDBC .......................... SUCCESS > [13:45 min] > 40946[INFO] > ------------------------------------------------------------------------ > 40947[INFO] BUILD SUCCESS > 40948[INFO] > ------------------------------------------------------------------------ > 40949[INFO] Total time: 14:07 min > 40950[INFO] Finished at: 2022-12-23T09:22:46Z {code} > > > The main problems that we see on testings are: > * Parameterized tests that setup database containers (this tends to be more > time consuming as another databases are added) > * Creation of containers by class (this is fine as we want test to be > independents, but the database could be setup 1time and cleaned at end of > each class) > * No easy way to extend a test to another database, at the end we copy a lot > of code > * A lot of code for create and populate tables for testing. We have > JdbcTestBase that use JdbcTestFixture to create a kind of 'Book Store', but > this is used in a lot of test without the use of all tables that are > implemented on store. > > -- This message was sent by Atlassian Jira (v8.20.10#820010)