Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
-
None
-
Reviewed
Description
Something tasks failed because of issues on NMs. For example, bad disk/network could cause reducer fetching failure and mappers need to be re-scheduled.
It will be very helpful to identify such issues if we could print host of failed tasks, which we can simply grep MR AM's log to see what happened.