0.9.2
版本发布时间: 2019-12-28 17:33:07
apache/linkis最新发布版本:1.6.0(2024-07-12 14:47:36)
Enhancement
- [Linkis-193] Deployment enhancements: Modify the install.sh script to support distributed and stand-alone installations.
- [Linkis-194] Dependency check: Check dependent linux commands before installing, and exit directly if no corresponding command is found.
- [Linkis-195] Service detection: Add the content to check whether all services start normally in the start-all.sh script.
- [Linkis-191] move hadoop dependencies from core/linkis-common to core/hadoop-common module.
- [Linkis-192] move httpclient dependencies from core/linkis-common to core/linkis-httpclient module.
- [Linkis-196] Modify the default port of microservices starting from 9100.
- [Linkis-197] Unified memory configuration and modify the default memory to 512m.
- [Linkis-198] ResourceManager request the rest interface of yarn to increase parameter configuration.
- [Linkis-199] unify the start script of microservices.
- [Linkis-200] to compile publicly dependent modules separately into a directory to reduce the size of the installation package.
- [Linkis-201] Add the hadoop / spark / hive configuration directory to the linkis configuration file by default.
- [Linkis-208 ] Add version check content in install.sh script.
- [Linkis-221 ] linkis-RM supports more hadoop versions.
- [Linkis-174 ] add hive postgresql metadata support.
Bug Fix
- [Linkis-175] update yarn's default queue name be changed from ide to default.
- [Linkis-202] deleted the hdfs dependency of errorCodemanager.
- [Linkis-205] Canonical sql comments.
- [Linkis-210] Fix bml hdfs permissions issue.
- [Linkis-218] remove the yum command in install.sh and start-all.sh.
- [Linkis-226] exclude Jackson packages in the metadata.
- [Linkis-229] exclude Jasper packages in public-module.
Credits
Last but not least, this release would not have been possible without the following contributors: allwefantasy: Added MLSQL engine. chenxi0599: Fix filesystem print error in log. wForget:Provides a release version of CDH5.7.6 for linkis. hj2016: Fixed a bug caused by maxEffectiveCapacity parameter not being present when Yarn gets the maximum resources for Capacity scheduling.. houjunxiong: Enhanced metadata support for meta database type of Hive to postgresql. Just-do-it-Fan: Enhanced ResourceManager module support for Yarn's Capacity scheduling. liangqilang: Fix the bug that NullPointException is thrown when downloading files from httpclient and add the function of streaming download result set, etc.. leisore: Fix start and stop script naming issues. nimuyuhan: Fix configuration file does not support Chinese bug. patinousward: Fix multiple bugs and add table building function in metadata module. zhanghaicheng1: Increase deployment and installation issue documentation.
增强
- [Linkis-193] 部署增强:脚本会自动识别是单机部署还是分布式部署,单机部署和启动时,不再使用SSH和SCP。
- [Linkis-194] 环境检查:在安装Linkis之前先检查依赖的所有Linux命令和环境,如果检查失败,则直接退出安装部署。
- [Linkis-195] 服务检测:在start-all脚本里面对服务是否正常启动进行检查,如果服务不正常则给出错误日志。
- [Linkis-191] 依赖优化:将hadoop相关工具类从公共core/common模块移到单独的core/hadoop-common模块,减少包冲突问题.
- [Linkis-192] 依赖优化:core/common模块不再依赖httpclient相关jar包,统一移到core/linkis-httpclient模块.
- [Linkis-196] 统一分配Linkis各个服务的端口,从9100开始分配.
- [Linkis-197] Linkis支持通过配置SERVER_HEAP_SIZE环境变量,统一指定所有微服务的堆内存大小.
- [Linkis-198] Linkis-RM除了通过读取yarn-site.xml自动获取Yarn的JMX URL的方式之外,新增一种通过参数配置的方式获取。
- [Linkis-199] 统一微服务的启动脚本,方便后续统一修改.
- [Linkis-200] 将公共依赖的模块放到public-module,减少包冲突和降低安装包大小.
- [Linkis-201] 默认将hadoop/hive/spark环境变量写到各个微服务的配置文件中.
- [Linkis-208 ] 在安装脚本中新增haoop/hive/spark版本兼容性检测,如果版本不兼容,会提醒用户是否选择继续安装.
- [Linkis-221 ] 增强Linkis-RM模块,使RM可以获取更多hadoop版本的Yarn队列信息.
- [Linkis-174 ] 增强metadata模块,使metadata支持访问postgresql类型的hive元数据库.
修复
- [Linkis-175] 将yarn的默认队列从ide修改为default.
- [Linkis-202] 移除Entrance错误码功能对hdfs的依赖,使精简版可以完全不依赖hadoop.
- [Linkis-205] SQL注释优化,在--后面加空格.
- [Linkis-210] 修复BML访问hdfs的权限问题.
- [Linkis-218] 从start和install脚本中移除yum命令,兼容更多的unix系统.
- [Linkis-226] 优化metadata模块,使引入的hive包,排除对Jackson包的依赖,降低包冲突风险.
- [Linkis-229] 排除公共依赖public-module中的Jasper包,降低包冲突风险.
- [Linkis-143] Linkis的配置文件增加对中文配置的支持.
百度云下载:
点击跳转百度云 提取码: iyya
贡献者
最后但是最重要的一点是,如果没有以下贡献者不可能发布此版本,在此感谢各位社区的贡献者! 排名不分先后(按字母排序): allwefantasy: 增加了MLSQL引擎. chenxi0599: 修复filesystem在日志打印错误问题. hj2016: 解决当Yarn为Capacity调度获取最大资源时,由于maxEffectiveCapacity参数不存在导致的bug. houjunxiong: 增强模块metadata对hive的元数据库类型为postgresql的支持. Just-do-it-Fan: 增加ResourceManager模块对Yarn的Capacity调度支持. leisore: 修复start和stop脚本命名问题. liangqilang: 修复httpclient下载文件抛NullPointException的bug和增加流传输下载结果集功能,等多个特性. nimuyuhan: 修复配置文件不支持中文bug. patinousward: 修复多个bug,并在metadata模块中增加建表功能. wForget:提供了linkis的CDH5.7.6可用的发布版本. zhanghaicheng1: 增加部署安装问题文档.
1、 wedatasphere-linkis-0.9.2-dist-CDH5.7.6-spark2.0-2.2.tar.gz 577.08MB
2、 wedatasphere-linkis-0.9.2-dist-spark2.0-2.2.tar.gz 524.23MB
3、 wedatasphere-linkis-0.9.2-dist-spark2.4.x.tar.gz 524.19MB