hadoop3版本的兼容性简单调研
目录:
1.与其他组件的兼容性版本
2.与hadoop2版本的客户端兼容性
一.与其他组件的兼容性版本
目前依赖hadoop的常用组件有spark、hive、hbase等.目前只调研hive和hbase这两个重要组件。
参考组件的官网发布信息,一般每个版本的发布信息会说明已经支持那个hadoop版本的底层服务,说明
已经经过官网比较系统的测试了。
1.1hive:
摘录官网的信息:
1 November 2018: release 3.1.1 available
This release works with Hadoop 3.x.y. You can look at the completeJIRA change log for this release.
30 July 2018: release 3.1.0 available
This release works with Hadoop 3.x.y. You can look at the completeJIRA change log for this release.
21 May 2018 : release 3.0.0 available
This release works with Hadoop 3.x.y. The on-disk layout of Acid tables has changed with this release. Any Acid table partition that had Update/Delete/Merge statement executed since the last Major compaction must execute Major compaction before upgrading to 3.0. No more Update/Delete/Merge may be executed against these tables since the start of Major compaction. Not following this may lead to data corruption. Tables/partitions that only contain results of Insert statements are fully compatible and don’t need to be compacted. You can look at the completeJIRA change log for this release.
可知,目前hive的三个罪行版本是支持hadoop3.x.y系列的,分别为:3.0.0、3.1.0、3.1.1。
1.2hbase
参考hbase官网的支持矩阵:
可知,Hbase2.0.x和Hbase2.1.x支持hadoop3.0.3+、hadoop3.1.1+;
Hbase2.1.x支持hadoop3.1.1+;
二.与hadoop2版本的客户端的兼容性
目前测试可以兼容hadoop2的客户端(rpc兼容),可以继续使用,但是一些新功能不能正常使用,且一些shell脚本发生了很大变化。
持续更新。。。。。