hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rui Li (JIRA)" <>
Subject [jira] [Commented] (HIVE-8991) Fix custom_input_output_format [Spark Branch]
Date Thu, 04 Dec 2014 08:08:12 GMT


Rui Li commented on HIVE-8991:

I looked a little more into this. It seems hive-exec is properly added to class path (as user
application jar in {{SparkSubmit}}) and class loader can load {{HiveIgnoreKeyTextOutputFormat}}:
2014-12-04 08:35:33,383 INFO  [stdout-redir-1]: client.SparkClientImpl (
- [Loaded from file:/home/hive/packaging/target/apache-hive-0.15.0-SNAPSHOT-bin/apache-hive-0.15.0-SNAPSHOT-bin/lib/hive-exec-0.15.0-SNAPSHOT.jar]
Nevertheless I still get the following error:
2014-12-04 08:32:26,681 INFO  [stderr-redir-1]: client.SparkClientImpl (
- java.lang.NoClassDefFoundError: org/apache/hadoop/hive/ql/io/HiveIgnoreKeyTextOutputFormat
Besides, the exception is thrown when we try to deserialize SparkWork in the job, which means
{{org.apache.hadoop.hive.ql.exec.spark.KryoSerializer}} has been loaded properly.
I'll do more debugging. Wondering if it's possible the error message is not accurate.

As for {{SparkSubmitDriverBootstrapper}} hanging issue, it's because it calls System.exit
in a shutdown hook which causes deadlock. It's been fixed in latest branch.

> Fix custom_input_output_format [Spark Branch]
> ---------------------------------------------
>                 Key: HIVE-8991
>                 URL:
>             Project: Hive
>          Issue Type: Sub-task
>          Components: Spark
>            Reporter: Rui Li
>            Assignee: Rui Li
>         Attachments: HIVE-8991.1-spark.patch
> After HIVE-8836, {{custom_input_output_format}} fails because of missing hive-it-util
in remote driver's class path.

This message was sent by Atlassian JIRA

View raw message