Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
Description
When Teragen is run multiple times with the same output directory, normally it should validate and fail.
But some cases it may continue and cause the exceptions which results failure in job later time.
I think the reason behind it is
org.apache.hadoop.examples.terasort.TeraOutputFormat.checkOutputSpecs(TeraOutputFormat.java) have issue, it permit the output already exists if it have only 1 kid and it's PARTITION_FILENAME