# Fix Ignore tenant conditions when refreshing a job Fix null key returned for cache operation when adding database Downgrade Guava to 31.1-jre to fix CDC pipeline Fix issue Failed to recyle database:database in auto-commit mode Fix role menu update issue Exception after submitting CDCSOURCE job to sync data from tables whose names start with digit(s) Add end time for jar sql # Optimization Dockerfile retains flink-table-planner-loader.jar Add an option in Flink on Kubernetes to control whether to include an OwnerReference Unified definition of Hadoop version Replace chinese punctuation with english punctuation marks and fix typo # Contributors @aiwenmo @cantiandashu @gaoyan1998 @gary-connear @leeoo @liguifa @safishhh @zackyoungh

# Feature - Support recover from the latest completed CheckPoint # Fix - Fix the flinksql history version switch button does not take effect - Fix the issue of the page constantly refreshing - Fix the issue where two Flink configuration items with the same prefix cause errors in Flink configuration parsing - Fix the issue where the form is cleared on the frontend when creating a new task - Sync the helm configuration with src/main/resources directory - Fix the issue where two Flink configuration items with the same prefix cause errors in Flink configuration parsing - Fix Array type in paimon query did not show the value - Fix metadata-hive MissingFormatArgumentException - Fix import errors in utility classes and Fix configuration issues related to frontend display - Fix comparison expression issue - Fix Paimon tinyint data can not preview in data source center - Fix an error in the comparison of Flink JAR task historical caches - Fix the WebSocket reconnection causes the page to flicker - Fix the initialization request on the login page is loaded multiple times # Optimize - Check if the cluster is disabled before starting the flink session - Optimize WebSocket for asynchronous sending - The WebSocket reconnection causes the page to flicker # Contributors @aiwenmo @cantiandashu @Jam804 @leeoo

# Feature - Support CALL statement - Flink Kubernetes opoerator supports ingress # Fix - Fix FLINK JAR submission - Fix deserialization exceptions caused by incorrect get methods in enum fields - Change k8s StringUtils import - Fix FlinkJar args global variable parsing - Fix variable can not throw error - Fix for debug task when target table contains '.' - Fix the lineage of task with variable - Fix postgresql when using concat - Fix for task lock strategy condition - Fix FlinkJar task lost info - Fix click to trigger savepoint error - Fix for reading 'root-exception' - Fix the errors when querying the data of numeric and date types in Paimon - Fix set statement is not effective in application mode - Fix for web socket session do not closed correctly # Optimize - Optimize the execution logic of the script - Optimize dolphin push info when enable English - Optimize docker image build - Optimization home page icon and job detail lineage under dark theme - Optimize lineage relationship chart display - Optimize e2e test - Add source url param in CDCSOURCE - The user-defined flink conf path overrides the flink conf path parameter # Document - Fix normal deploy doc error - Quick experience doc update # Contributors @aiwenmo @gaoyan1998 @Jam804 @liguifa @MactavishCui @zackyoungh @Zzm0809

# New features - Added domestic npm mirror sources (supported by profile mode) - Support Flink-1.20 - Support FlinkCDC 3.2.0 - Added built-in flink history server, which greatly reduces the problem of incorrect task status acquisition - Added some page prompts for roles and permissions - Added loading effects when displaying blood relationship acquisition - Implemented job import and export - Support mock test flinksql and CDCSOURCE task - FlinkSQL Studio supports real-time update task status - Add a boot initialization page for the first deployment # Fixes - Fix the problem that show databases cannot be executed normally - Fix some json serialization problems - Fix the problem of SQL Cli under Flink1.19 - Fix the problem that the task port is unavailable in k8s mode - Fix some problems in the git project function (default value, drag and drop sorting) - Fix the SavePoint path logic in kubernetes-operator mode and adjust the configuration method of Flink configuration acquisition - Fix the problem that the page is unresponsive for a long time when too many tasks are opened on the data development page - Fix git Problems in the project build process - Fix the thumbnail display problem in the code editor - Fix the sql automatic initialization problem of pg - Fix the problem of submitting tasks in local mode - Fix the data type conversion problem when synchronizing the entire database of Oracle - Fix the exception caused by no instance when clicking on the workbench - Fix the syntax error problem when querying data in postgres - Fix the problem that flyway cannot support mysql5.7 - Fix the problem of incorrect primary key column acquisition in Oracle data source type - Fix the problem that the task list sorting does not take effect - Fix the problem of repeated refresh of Git project page - Fix the null pointer exception that may exist in DingTalk alarm - Fix the problem of proxy Flink address - Fix the problem that the table structure in the catalog cannot be displayed normally # Optimization - Optimize the support of https protocol when obtaining task details - Optimize the workbench page layout - Delete the unique index of the dinky_cluster table to solve the problem in yarn/k8s Unique index conflict problem when high availability is enabled - Optimize some Mapper queries - Optimize the type of declared attributes of the git project backend class - Delete the prompt text in UDF registration management - Optimize some page layouts to make them more user-friendly on small screens - Optimize the version update logic to solve the cache problem caused by the upgrade (automatic version comparison implementation) - Optimize the virtual scrolling failure to take effect when there are too many nodes in the data source details list tree, resulting in slow page rendering - Optimize the login page to solve the problem of excessive resource jams - Optimize the application startup speed - Optimize the cluster started from the cluster configuration to manual registration - Optimize the display layout of the system configuration description that is too long - Optimize the overall layout and rendering efficiency of the flink task operator graph in the operation and maintenance center # Refactoring - Refactor the user.dir acquisition method to avoid the problem of incorrectly obtaining the project root path in different deployment environments - Refactor SSE to websocket - Refactor the request method for obtaining task monitoring data - Remove hutool-json to improve data conversion efficiency # Documentation - Added documentation for integrating Dinky into Datasophon - Added documentation for using SQL Cli - Optimized the document footer - Added documentation for role binding tenants # CI/CD - Added optional version items for the github bug template - Removed the domestic warehouse proxy configuration when deploying documents

# New features - Added domestic npm mirror sources (supported by profile mode) - Support Flink-1.20 - Support FlinkCDC 3.2.0 - Added built-in flink history server, which greatly reduces the problem of incorrect task status acquisition - Added some page prompts for roles and permissions - Added loading effects when displaying blood relationship acquisition - Implemented job import and export - Support mock test flinksql and CDCSOURCE task - FlinkSQL Studio supports real-time update task status - Add a boot initialization page for the first deployment # Fixes - Fix the problem that show databases cannot be executed normally - Fix some json serialization problems - Fix the problem of SQL Cli under Flink1.19 - Fix the problem that the task port is unavailable in k8s mode - Fix some problems in the git project function (default value, drag and drop sorting) - Fix the SavePoint path logic in kubernetes-operator mode and adjust the configuration method of Flink configuration acquisition - Fix the problem that the page is unresponsive for a long time when too many tasks are opened on the data development page - Fix git Problems in the project build process - Fix the thumbnail display problem in the code editor - Fix the sql automatic initialization problem of pg - Fix the problem of submitting tasks in local mode - Fix the data type conversion problem when synchronizing the entire database of Oracle - Fix the exception caused by no instance when clicking on the workbench - Fix the syntax error problem when querying data in postgres - Fix the problem that flyway cannot support mysql5.7 - Fix the problem of incorrect primary key column acquisition in Oracle data source type - Fix the problem that the task list sorting does not take effect - Fix the problem of repeated refresh of Git project page - Fix the null pointer exception that may exist in DingTalk alarm - Fix the problem of proxy Flink address - Fix the problem that the table structure in the catalog cannot be displayed normally # Optimization - Optimize the support of https protocol when obtaining task details - Optimize the workbench page layout - Delete the unique index of the dinky_cluster table to solve the problem in yarn/k8s Unique index conflict problem when high availability is enabled - Optimize some Mapper queries - Optimize the type of declared attributes of the git project backend class - Delete the prompt text in UDF registration management - Optimize some page layouts to make them more user-friendly on small screens - Optimize the version update logic to solve the cache problem caused by the upgrade (automatic version comparison implementation) - Optimize the virtual scrolling failure to take effect when there are too many nodes in the data source details list tree, resulting in slow page rendering - Optimize the login page to solve the problem of excessive resource jams - Optimize the application startup speed - Optimize the cluster started from the cluster configuration to manual registration - Optimize the display layout of the system configuration description that is too long - Optimize the overall layout and rendering efficiency of the flink task operator graph in the operation and maintenance center # Refactoring - Refactor the user.dir acquisition method to avoid the problem of incorrectly obtaining the project root path in different deployment environments - Refactor SSE to websocket - Refactor the request method for obtaining task monitoring data - Remove hutool-json to improve data conversion efficiency # Documentation - Added documentation for integrating Dinky into Datasophon - Added documentation for using SQL Cli - Optimized the document footer - Added documentation for role binding tenants # CI/CD - Added optional version items for the github bug template - Removed the domestic warehouse proxy configuration when deploying documents

# New features - Added domestic npm mirror sources (supported by profile mode) - Support Flink-1.20 - Support FlinkCDC 3.2.0 - Added built-in flink history server, which greatly reduces the problem of incorrect task status acquisition - Added some page prompts for roles and permissions - Added loading effects when displaying blood relationship acquisition - Implemented job import and export - Support mock test flinksql and CDCSOURCE task - FlinkSQL Studio supports real-time update task status - Add a boot initialization page for the first deployment # Fixes - Fix the problem that show databases cannot be executed normally - Fix some json serialization problems - Fix the problem of SQL Cli under Flink1.19 - Fix the problem that the task port is unavailable in k8s mode - Fix some problems in the git project function (default value, drag and drop sorting) - Fix the SavePoint path logic in kubernetes-operator mode and adjust the configuration method of Flink configuration acquisition - Fix the problem that the page is unresponsive for a long time when too many tasks are opened on the data development page - Fix git Problems in the project build process - Fix the thumbnail display problem in the code editor - Fix the sql automatic initialization problem of pg - Fix the problem of submitting tasks in local mode - Fix the data type conversion problem when synchronizing the entire database of Oracle - Fix the exception caused by no instance when clicking on the workbench - Fix the syntax error problem when querying data in postgres - Fix the problem that flyway cannot support mysql5.7 - Fix the problem of incorrect primary key column acquisition in Oracle data source type - Fix the problem that the task list sorting does not take effect - Fix the problem of repeated refresh of Git project page - Fix the null pointer exception that may exist in DingTalk alarm - Fix the problem of proxy Flink address - Fix the problem that the table structure in the catalog cannot be displayed normally # Optimization - Optimize the support of https protocol when obtaining task details - Optimize the workbench page layout - Delete the unique index of the dinky_cluster table to solve the problem in yarn/k8s Unique index conflict problem when high availability is enabled - Optimize some Mapper queries - Optimize the type of declared attributes of the git project backend class - Delete the prompt text in UDF registration management - Optimize some page layouts to make them more user-friendly on small screens - Optimize the version update logic to solve the cache problem caused by the upgrade (automatic version comparison implementation) - Optimize the virtual scrolling failure to take effect when there are too many nodes in the data source details list tree, resulting in slow page rendering - Optimize the login page to solve the problem of excessive resource jams - Optimize the application startup speed - Optimize the cluster started from the cluster configuration to manual registration - Optimize the display layout of the system configuration description that is too long - Optimize the overall layout and rendering efficiency of the flink task operator graph in the operation and maintenance center # Refactoring - Refactor the user.dir acquisition method to avoid the problem of incorrectly obtaining the project root path in different deployment environments - Refactor SSE to websocket - Refactor the request method for obtaining task monitoring data - Remove hutool-json to improve data conversion efficiency # Documentation - Added documentation for integrating Dinky into Datasophon - Added documentation for using SQL Cli - Optimized the document footer - Added documentation for role binding tenants # CI/CD - Added optional version items for the github bug template - Removed the domestic warehouse proxy configuration when deploying documents

## Dinky-1.0.3 Release Note ### Upgrade Instructions > 1.0.3 is a bug fix version, no table structure changes, no additional SQL scripts need to be executed during upgrade, just overwrite and install, pay attention to the modification of configuration files and the placement of dependencies > About SCALA version: The release uses Scala-2.12. If your environment must use Scala-2.11, please compile it yourself, please refer to [Compile and Deploy](../docs/next/deploy_guide/compile_deploy), change the profile scala-2.12 to scala-2.11 ### New Features - Added the function of manually killing the process after the task is stuck during operation ### Fixes - Fix the problem that the Yarn Application mode cannot execute tasks in Flink 1.19 support - Fix the problem of start and stop scripts, adapt to the GC parameters of jdk 11 - Fix UDF The class cannot be found after publishing - Fixed the priority problem that the set function cannot cover in the Application task SQL ### Optimization - Optimize the Dinky service to cause the CPU load to be too high and the thread not to be released during monitoring - Optimize the Dinky monitoring configuration, according to the `Configuration Center->Global Configuration->Metrics Configuration->**Dinky JVM Monitor Switch**` switch to control whether to enable Flink task monitoring - Optimize the data type conversion logic of Oracle whole database synchronization - Optimize the front-end rendering performance and display effect of monitoring data

## Dinky-1.0.2 ### Upgrade Instructions > - 1.0.2 is a BUG repair version with table structure/data changes, please execute DINKY_HOME/sql/upgrade/1.0.2_schema/data source type/dinky_dml.sql ### New Feature - Adapt to various Rest SvcTypes in KubernetsApplicationOperator mode and modify JobId to obtain judgment logic - Added SSE heartbeat mechanism - Added the function of automatically retrieving the latest highly available JobManager address (currently implemented in Yarn; not yet implemented in K8s) - Added the function of clearing logs in the console during data development - Support Flink1.19 - Add task group related configuration when pushing to Apache DolphinScheduler - Added a user designated by the user to submit YarnApplication tasks - The startup script adds GC related startup parameters and supports configuring the DINKY_HOME environment variable - Implement FlinkSQL configuration item in cluster configuration to support RS protocol (Yarn mode only) ### Fix - Fixed the problem of global variables not being recognized in YarnApplication mode, and reconstructed the YarnApplication submission method - Fixed the problem of data source heartbeat detection feedback error - Fix the possible 404 issue in front-end route jump - Fixed the issue of incorrect error prompt when global variable does not exist - Fixed the issue of cursor movement and flickering in the editor during front-end data development - Fixed the path error problem in the docker file of DockerfileDinkyFlink - Fixed the problem of unrecognized configuration Python options - Fixed null pointer exception in role user list - Fixed some issues when submitting K8s tasks - Fixed Oracle's Time type conversion problem when synchronizing the entire database - Fixed the problem that k8s pod template cannot be parsed correctly - Fixed the issue where SPI failed to load CodeGeneratorImpl - Fixed an issue where numeric columns declared with UNSIGNED / ZEROFILL keywords would cause parsing mismatches - Fixed the issue where the status of batch tasks is still unknown after completion - Fixed some unsafe interfaces that can be accessed without login authentication - Fixed the problem of unknown status in Pre-Job mode - Fixed the problem of retrieving multiple job instances due to duplicate Jid - Fixed the problem that the user list cannot be searched using worknum - Fixed the problem that the query data button on the right side of the result Tag page cannot be correctly rendered when querying data. - Fixed issues with print table syntax - Fixed the problem that the resource list cannot be refreshed after adding or modifying it - Fixed the issue of incorrect console rolling update task status for data development - Fixed the problem of occasional packaging failure - Fixed problems when building Git projects ### Optimization - Optimize start and stop scripts - Optimize the problem of partial overflow of the global configuration page - Tips for optimizing UDF management - Optimize the user experience of the operation and maintenance center list page and support sorting by time - Optimize the warehouse address of default data in Git projects - Optimize flink jar task submission to support batch tasks - Optimize the problem that the right-click menu cannot be clicked when it overflows the visual area. - Optimize the primary key key of the list component of the operation and maintenance center - When modifying tasks, the modifiable template is optimized to an unmodifiable template. - Optimize the display method and type of cluster configuration - Optimize the logic of deleting clusters in K8s mode - Fixed the problem that the cluster is not automatically released in Application mode - Remove the logic of using Paimon for data source caching and change it to the default memory cache, which can be configured as redis cache - Removed the automatic pop-up of Console when switching tasks - Optimize the rendering logic of resource management. The resource management function cannot be used when resources are not turned on. - Optimize the detection logic of login status - Optimize login page feedback prompts - Removed some useless code on the front end - Optimize the problem that when the entire library is synchronized to build operator graphs multiple times, the order of the operator graphs is inconsistent, resulting in the inability to recover from the savepoint. - Some tips for optimizing resource allocation - Optimize and improve the replication function of the resource center, supporting all reference scenarios within Dinky currently ### Safety - Exclude some high-risk jmx exposed endpoints ### Document - Optimize expression variable expansion documentation - Optimize some practical documents for synchronization of the entire database - Add JDBC FAQ about tinyint type - Added a carousel image on the home page of the official document website - Fixed the description problem of resource configuration in document global configuration - Added documents related to environment configuration in global configuration - Delete some configuration items of Flink configuration in the global configuration - Added document configuration description for alarm type email