Retain data to debug failed tasks for MapReduce jobs

You can configure IBM® Spectrum Symphony Advanced Edition to retain job and task working directories, footprint files, map task output intermediate data files, distributed cache files, and core dump files for MapReduce jobs. Use this saved data to help debug and troubleshoot your MapReduce workload. For example, if JVM crashes while running jobs, use the saved data to view details and diagnose your system. This data is useful for your own diagnosis, or for providing to IBM Support, if required.

To configure the MapReduce framework to retain data, either edit the pmr-site.xml configuration file, or run the mrsh utility using the -D option.