[PIG-4101] Lower tez.am.task.max.failed.attempts to 2 from 4 in Tez mini cluster

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.14.0
• Component/s: tez
• Assignee: Cheolsoo Park
• Reporter: Cheolsoo Park
• Created: 04/Aug/14 18:28
• Updated: 21/Nov/14 05:59
• Resolved: 04/Aug/14 19:31

Description

Currently, Tez mini cluster tests are slow partly because negative tests are slow. The reason is that Tez retries 4 times before marking the job failed.

I suggest we should lower the max failed attempts to 2 so that negative tests will fail faster. For example, TestAccumulator runs 30 secs faster with this change-

tez.am.task.max.failed.attempts=2: Tests run: 23, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 276.269 sec
tez.am.task.max.failed.attempts=4: Tests run: 23, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 302.772 sec

I am not lowering it to 1 for transient errors.