Hi Witgo,
Thanks a lot for your reply.
In my second test, the user features and product features were loaded
from the file system directly,which means I did not use ALS here, and this
problem happened at the loading data stage. The way I am asking the question
was a little bit miss leading, this problem might not be ALS specific.
As far as I had discovered, it might related to the usage of master
memory, since I could solve this problem by increase the master memory size,
or reduce the number of partitions. But I am still waiting for a better
explanation :)
Thank you again.
Sue
--
View this message in context:
http://apache-spark-developers-list.1001551.n3.nabble.com/MLlib-ALS-Errors-communicating-with-MapOutputTracker-tp6740p6758.html
Sent from the Apache Spark Developers List mailing list archive at Nabble.com.
