[
https://issues.apache.org/jira/browse/HUDI-4415?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17570714#comment-17570714
]
XiaoyuGeng edited comment on HUDI-4415 at 7/25/22 7:54 AM:
-----------------------------------------------------------
[~rmahindra] If the problem has high priority according to the feedback, i will
raise a pr to solve it. By now, i get a simple implementation as a workaround
was (Author: minihippo):
[~rmahindra] If the problem is high priority according to the feedback, i will
raise a pr to solve it. By now, i get a simple implementation as a workaround
> Support spark writer running on thrift server
> ---------------------------------------------
>
> Key: HUDI-4415
> URL: https://issues.apache.org/jira/browse/HUDI-4415
> Project: Apache Hudi
> Issue Type: New Feature
> Components: spark-sql
> Reporter: XiaoyuGeng
> Priority: Major
>
> Spark thrift server manages a sql by running with the thread local spark
> session. It identifies the jobs belong to the sql by setting the job group
> Id. If user wants to cancel the sql running, thrift server will call
> session.cancelJobGroup. However, the implementation of hudi spark writer uses
> different job group Ids to distinguish the different proccesing part of the
> writing progress, like workprofile, compaction, rollback and so on. There is
> an Incompatibility
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
