Details
-
Task
-
Status: Closed
-
Major
-
Resolution: Fixed
-
5.1.0
-
None
Description
When using ${coord:latest(n)} coordinator EL function inside an input dataset dependency, it's often the case that more information is needed how many HDFS URIs are being checked for each <data-in/>.
Right now we don't have this information. While debugging and fine tuning parameters like dataset frequency, initial-instance, and data-in instance, it would be very useful to know how many HDFS roundtrips are issues by the current settings CoordELFunctions#coord_latestRange_sync() and CoordELFunctions#coord_futureRange_sync() having called DFSClient#exists(). We need appropriate logging there.
Attachments
Attachments
Issue Links
- blocks
-
OOZIE-3387 Optimize coordinator data input dependency search
- Open
- links to