Open topic with navigation
When a MapReduce job uses functionality that is not supported by the DMExpress Hadoop Sort plugin, DMExpress will be unable to accelerate the sort performance, and the job will fall back to the native Hadoop sort. This happens seamlessly and does not affect the successful running of the job.
You can confirm that fallback has occurred by checking the Hadoop job logs.
To confirm whether fallback has occurred, check the Hadoop logs for the MapReduce job.
2013-03-28 11:08:06,328 INFO com.syncsort.dmexpress.hadoop.DMXMapOutputCollector: DMExpress sort plugin is not supported, fallback to default sort: java.lang.UnsupportedOperationException: Key class com.mycompany.io.MyCustomWritable is not supported
For information on finding the Hadoop job logs, see Finding DMExpress Hadoop Logs on MRv1 or Finding DMExpress Hadoop Logs on YARN (MRv2).
Copyright © 2016 Syncsort All rights reserved.