You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "sydt (JIRA)" <ji...@apache.org> on 2017/11/13 11:33:00 UTC
[jira] [Commented] (SPARK-22506) Spark thrift server can not
impersonate user in kerberos
[ https://issues.apache.org/jira/browse/SPARK-22506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16249422#comment-16249422 ]
sydt commented on SPARK-22506:
------------------------------
From logs, executors does not have permission to access hdfs file which means executor does not have blocktoken to access datanode.
> Spark thrift server can not impersonate user in kerberos
> ---------------------------------------------------------
>
> Key: SPARK-22506
> URL: https://issues.apache.org/jira/browse/SPARK-22506
> Project: Spark
> Issue Type: Improvement
> Components: Deploy
> Affects Versions: 2.2.0
> Reporter: sydt
> Attachments: screenshot-1.png
>
>
> Spark thrift server can not impersonate user in kerberos environment.
> I launch spark thrift server in* yarn-client *mode by user *hive* ,which is allowed to impersonate other user.
> User* jt_jzyx_project7* submit sql statement to query its own table located in hdfs catalog: /user/jt_jzyx_project7, and happened errors:
> Permission denied: *user=hive*, access=EXECUTE, inode=*"/user/jt_jzyx_project7*":hdfs:jt_jzyx_project7:drwxrwx---:user:g_dcpt_project1:rwx,group::rwx
> obviously, spark thrift server didn't proxy user: jt_jzyx_project7 .
> And this happened task stage, which means it pass the hive authorization.
> !screenshot-1.png!
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org