v1.1.0

版本发布时间: 2024-07-30 22:03:38

DataLinkDC/dinky最新发布版本:v1.1.0(2024-07-30 22:03:38)

Dinky-1.1.0 Release Note

v1.1.0 supports the automatic schema upgrade framework (Flyway), using the table structure/data up to v1.0.2 as the default base version. If your version is not at v1.0.2+, you must first upgrade to the table structure of v1.0.2 according to the official upgrade tutorial. If your version is v1.0.2+, you can directly upgrade, and the program will automatically execute without affecting historical data. If you are deploying from scratch, please ignore this matter.
Due to the contribution of flink-cdc to the Apache Foundation, the package name of the new version will change, and it is not possible to make compatibility changes. In versions dinky-v1.1.0 and above, dinky will use new package name dependencies, which requires your flink-cdc dependencies to be upgraded to flink-cdc v3.1+, otherwise it will not work.
Remove the distinction of Scala version when packaging, only develop with Scala-2.12, and no longer support Scala-1.11.x.

Added Flyway schema upgrade framework.
Task directory supports flexible sorting.
Implemented task-level permission control and supports different permission control strategies.
Optimized the automatic addition of administrator user association when adding tenants.
Added the function to directly kill the process in the case of task submission deadlock.
Support k8s deployment of dinky.
Implement data preview.
New support for UDF injection configuration in data development.
Added whole library synchronization function (cdcsource) sink end table name mapping, regular matching modification mapping.
Added Dashboard page.
Added Paimon data source type.
Added SQL-Cli.

Modified the issue with k8s's account.name value and added the problem of Conf initialization when deleting a cluster.
Fixed the issue of flink-cdc losing SQL in application mode.
Fixed the issue where the task creation time was not reset when copying tasks.
Fixed the task list positioning problem.
Solved the problem of user-defined classes in user Jars not being compiled when submitting Jar tasks.
Fixed the incorrect alarm information in the enterprise WeChat-app mode.
Fixed the problem of flink-1.19 not being able to submit tasks.
Fixed the startup script not supporting jdk11.
Fixed the problem of cluster instances not being deleted.
Fixed the problem of UDF not finding the class in Flink SQL tasks.
Fixed the problem of the data development page not updating the state when the size changes.
Fixed the problem of not being able to get the latest high availability address defined in custom configuration.
Fixed the problem of not recognizing the manual configuration of rest.address and rest.port.

Optimized the prompt words in resource configuration.
Optimized the DDL generation logic of the MySQL data source type.
Optimized some front-end dependencies and front-end prompt information.
Optimized the copy path function of the resource center, supporting multiple application scenarios within dinky.
Optimized the monitoring function, using the monitoring function switch in dinky's configuration center to control all monitoring within dinky.
Optimized some front-end judgment logic.

Moved the alarm rules to the alarm route under the registration center.
Removed paimon as the storage monitoring medium, changed to sqllite, and do not strongly depend on the hadoop-uber package (except in the Hadoop environment), and support periodic cleaning.
Restructured the monitoring page, removing some built-in service monitoring.

Added documentation for deploying dinky on k8s.
Optimized the Docker deployment documentation.
Added documentation related to whole library synchronization function (cdcsource) sink end table name mapping.