[
https://issues.apache.org/jira/browse/HIVE-12160?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14971247#comment-14971247
]
Aihua Xu commented on HIVE-12160:
---------------------------------
Thanks for also testing that out.
I'm also investigating. Will try to provide a patch. Seems we should relogin
with the service principal in the child process since otherwise, it will run as
the user of the process "hive" rather than the service principal, then it will
fail to authenticate against Hbase or HDFS.
> Hbase table query execution fails in secured cluster when
> hive.exec.mode.local.auto is set to true
> --------------------------------------------------------------------------------------------------
>
> Key: HIVE-12160
> URL: https://issues.apache.org/jira/browse/HIVE-12160
> Project: Hive
> Issue Type: Bug
> Components: Security
> Affects Versions: 1.1.0
> Reporter: Aihua Xu
>
> In a secured cluster with kerberos, a simple query like {{select count(*)
> from hbase_table; }} will fail with the following exception when
> hive.exec.mode.local.auto is set to true.
> {noformat}
> Error: Error while processing statement: FAILED: Execution Error, return code
> 134 from org.apache.hadoop.hive.ql.exec.mr.MapRedTask (state=08S01,code=134)
> {noformat}
> There is another scenario which may be caused by the same reason.
> Set hive.auto.convert.join to true, the join query {{select * from hbase_t1
> join hbase_t2 on hbase_t1.id = hbase_t2.id;}} also fails with the following
> exception:
> {noformat}
> Error while processing statement: FAILED: Execution Error, return code 2 from
> org.apache.hadoop.hive.ql.exec.mr.MapredLocalTask (state=08S01,code=2)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
