[HIVE-2766] Hive freezes and takes 100% cpu when missing an HDFS block

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: CLI, Query Processor

Environment

Reproduced on Cloudera 3

• Assignee: Unassigned
• Reporter: Maxime Brugidou
• Created: 31/Jan/12 09:13
• Updated: 31/Jan/12 09:13

Description

When a block is corrupt or missing from HDFS (due to the failure of multiple nodes), Hive jobs hang mysteriously at the beginning (without starting the M/R job):

Launching Job 2 out of 4
Number of reduce tasks not specified. Estimated from input data size: 2
In order to change the average load for a reducer (in bytes):
  set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
  set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
  set mapred.reduce.tasks=<number>

The java process takes 100% of 1 CPU, and never stops, i had to kill manually all my hive jobs.

It would be nice if the processes could fail with an appropriate exception rather than hanging.