[HIVE-9882] Add jar/file doesn't work with yarn-cluster mode [Spark Branch]

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: spark-branch
• Fix Version/s: 1.2.0
• Component/s: Hive, Spark
• Assignee: Rui Li
• Reporter: Xiaomin Zhang
• Created: 06/Mar/15 02:25
• Updated: 04/Nov/15 09:43
• Resolved: 10/Mar/15 03:07

Description

It seems current fix for HIVE-9425 only uploads the Jar/Files to HDFS, however, they are not accessible by the Driver/Executor.
I found below in the AM log:

15/02/26 15:10:36 INFO Configuration.deprecation: mapred.min.split.size is deprecated. Instead, use mapreduce.input.fileinputformat.split.minsize
15/02/26 15:10:36 INFO client.SparkClientUtilities: Added jar[file:/data/hadoop-devel/data/nm/usercache/user/appcache/application_1424933948132_0002/container_1424933948132_0002_01_000001/hdfs:/localhost:8020/tmp/hive/user/47040bca-1da4-49b6-b2c7-69be9bc92855/hive-exec-1.2.0-SNAPSHOT.jar] to classpath.
15/02/26 15:10:36 INFO client.SparkClientUtilities: Added jar[file:/data/hadoop-devel/data/nm/usercache/user/appcache/application_1424933948132_0002/container_1424933948132_0002_01_000001/hdfs:/localhost:8020/tmp/hive/user/47040bca-1da4-49b6-b2c7-69be9bc92855/opennlp-maxent-3.0.3.jar] to classpath.
15/02/26 15:10:36 INFO client.SparkClientUtilities: Added jar[file:/data/hadoop-devel/data/nm/usercache/user/appcache/application_1424933948132_0002/container_1424933948132_0002_01_000001/hdfs:/localhost:8020/tmp/hive/user/47040bca-1da4-49b6-b2c7-69be9bc92855/bigbenchqueriesmr.jar] to classpath.
15/02/26 15:10:36 INFO client.SparkClientUtilities: Added jar[file:/data/hadoop-devel/data/nm/usercache/user/appcache/application_1424933948132_0002/container_1424933948132_0002_01_000001/hdfs:/localhost:8020/tmp/hive/user/47040bca-1da4-49b6-b2c7-69be9bc92855/opennlp-tools-1.5.3.jar] to classpath.
15/02/26 15:10:36 INFO client.SparkClientUtilities: Added jar[file:/data/hadoop-devel/data/nm/usercache/user/appcache/application_1424933948132_0002/container_1424933948132_0002_01_000001/hdfs:/localhost:8020/tmp/hive/user/47040bca-1da4-49b6-b2c7-69be9bc92855/jcl-over-slf4j-1.7.5.jar] to classpath.
15/02/26 15:10:36 INFO client.RemoteDriver: Failed to run job 6886df05-f430-456c-a0ff-c7621db712d6
org.apache.hive.com.esotericsoftware.kryo.KryoException: Unable to find class: de.bankmark.bigbench.queries.q10.SentimentUDF 

As above shows, the file path which was attempted to add to Classpath is invalid, so actually all uploaded Jars/Files are still not available for Driver/Executor.

Issue Links

• is related to: HIVE-9425 Add jar/file doesn't work with yarn-cluster mode [Spark Branch]
• links to: RB entry