Hello, Thanks for your reply.
Yes, Its netflix dataset. And when I get no space on device, my ‘/mnt’ directory gets filled up. I checked. /usr/lib/spark/bin/spark-submit --deploy-mode cluster --master yarn --class org.apache.spark.examples.mllib.MovieLensALS --jars /usr/lib/spark/examples/jars/scopt_2.11-3.3.0.jar /usr/lib/spark/examples/jars/spark-examples_2.11-2.0.0.jar --rank 32 --numIterations 100 --kryo s3://dataset_netflix When I run above command, I get following error Job aborted due to stage failure: Task 221 in stage 53.0 failed 4 times, most recent failure: Lost task 221.3 in stage 53.0 (TID 9817, ): java.io.FileNotFoundException: /mnt/yarn/usercache/hadoop/appcache/application_1473786456609_0042/blockmgr-045c2dec-7765-4954-9c9a-c7452f7bd3b7/08/shuffle_168_221_0.data.b17d39a6-4d3c-4198-9e25-e19ca2b4d368 (No space left on device) I think I should not need to increase the space on device, as data is not that big. So, is there any way, I can setup parameters so that it does not use much disk space. I don’t know much about tuning parameters. It will be great if anyone can help me with this. Thanks, Roshani > On Sep 16, 2016, at 9:18 AM, Sean Owen <so...@cloudera.com> wrote: > >