[SPARK-18626] Concurrent write to table fails from spark

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Incomplete
• Affects Version/s: 1.5.0
• Component/s: SQL
• Labels: bug bulk-closed
• Assignee: Unassigned
• Reporter: Thomas Sebastian
• Created: 29/Nov/16 12:10
• Updated: 27/May/20 06:57
• Resolved: 21/May/19 04:15

Description

When a spark job is submitted twice, to execute concurrently using spark summit, both the jobs are failing, not allowing concurrent write(Append) to Hive tables.

ERROR InsertIntoHadoopFsRelation: Aborting job.
java.io.IOException: Failed to rename FileStatus

{path=hdfs://nameservice1/user/hive/warehouse/aaa.db/table1/_temporary/0/task_201611210639_0017_m_000050/part-r-00050-00e873af-e3ab-4730-881f-e8a1b22077e0.gz.parquet; isDirectory=false; length=492; replication=3; blocksize=134217728; modification_time=1479728364366; access_time=1479728364062; owner=name; group=hive; permission=rw-rw-r--; isSymlink=false}

to hdfs://nameservice1/user/hive/warehouse/aaa.db/table1t/part-r-00050-00e873af-e3ab-4730-881f-e8a1b22077e0.gz.parquet
at org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter.mergePaths(FileOutputCommitter.java:371)
at org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter.mergePaths(FileOutputCommitter.java:384)
at org.apache.hadoop.mapreduce.lib.output.FileOutputCommitter.commitJob(FileOutputCommitter.java:326)
at parquet.hadoop.ParquetOutputCommitter.commitJob(ParquetOutputCommitter.java:46)