[
https://issues.apache.org/jira/browse/IMPALA-11232?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Fang-Yu Rao resolved IMPALA-11232.
----------------------------------
Resolution: Fixed
Resolve the issue since the fix has been merged.
> Consider not adding some jars when starting HMS
> -----------------------------------------------
>
> Key: IMPALA-11232
> URL: https://issues.apache.org/jira/browse/IMPALA-11232
> Project: IMPALA
> Issue Type: Test
> Components: Infrastructure
> Affects Versions: Impala 4.0.0
> Reporter: Fang-Yu Rao
> Assignee: Fang-Yu Rao
> Priority: Minor
>
> Currently when we execute {{run-hive-server.sh}} with the "{{-with_ranger}}"
> flag, we add all the jar files under the folder of
> {{$RANGER_HOME/ews/webapp/WEB-INF/lib}} to '{{HADOOP_CLASSPATH}}' before we
> call the shell script that starts the Hive service, i.e., {{bin/hive}}.
> Recall that {{bin/hive}} would attempt to execute '{{hadoop version}}' to
> determine the version of the underlying Hadoop distribution before starting
> the Hive Metastore and the HiveServer2. By default, this Hadoop command under
> the covers would add every path on '{{HADOOP_CLASSPATH}}' to its own
> environment variable of '{{CLASSPATH}}'. When there are too many paths added
> to '{{CLASSPATH}}', the Hadoop command would fail with the error message of
> "{{Argument list too long.}}" and hence the Hive service could not be started.
> We execute {{run-hive-server.sh}} with the "{{-with_ranger}}" flag in order
> to run the end-to-end test of {{test_hive_with_ranger_setup}}. We should
> consider not adding some (or all) of the jar files to '{{HADOOP_CLASSPATH}}'
> from the folder described if some (or all) of them are not required to run
> the end-to-end test so as to avoid the error mentioned above.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
