[CARBONDATA-1754] Carbon1.3.0 Concurrent Insert overwrite-Compaction: Compaction job fails at run time if insert overwrite job is running concurrentlyInsert overwrite

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.3.0
• Component/s: data-load
• Labels: dfx
• Environment: 3 Node ant cluster
• Assignee: Jatin
• Reporter: Ajeet Rai
• Created: 17/Nov/17 08:58
• Updated: 08/Jan/18 10:41

Description

Carbon1.3.0 Concurrent Insert overwrite-Compaction: Compaction job fails at run time if insert overwrite job is running concurrently.

Steps:
1: Create a table
2: Start three load one by one
3: After load is completed, start insert overwrite and minor compaction concurrently from two different session
4: observe that both jobs are are running
5: Observe that Insert overwrite job is success but after that compaction fails with below exception:

ERROR

[pool-23-thread-49]

Error running hive query:

org.apache.spark.sql.hive.thriftserver.SparkExecuteStatementOperation$$anon$1$$anon$2.run(SparkExecuteStatementOperation.scala:167) org.apache.hive.service.cli.HiveSQLException: java.lang.RuntimeException: Compaction failed. Please check logs for more info. Exception in compaction java.lang.Exception: Compaction failed to update metadata for table ajeet.flow_carbon_new999

7: Ideally compaction job should give error in start with message that insert overwrite in progress.

Issue Links

• links to: GitHub Pull Request #1711