Are you using the default Java object serialization, or have you tried Kryo yet? If you haven't tried Kryo please do and let me know how much it impacts the serialization size. (I know its more efficient, I'm curious to know how much more efficient, and I'm being lazy - I don't have ~6K 500MB files on hand.)
You can saveAsObjectFile on maybe a take(1) from an RDD and examine the serialized output to see if maybe a much larger graph than you expect is being output. On Fri Jan 30 2015 at 3:47:31 PM ankits <ankitso...@gmail.com> wrote: > This is on spark 1.2 > > I am loading ~6k parquet files, roughly 500 MB each into a schemaRDD, and > calling count() on it. > > After loading about 2705 tasks (there is one per file), the job crashes > with > this error: > Total size of serialized results of 2705 tasks (1024.0 MB) is bigger than > spark.driver.maxResultSize (1024.0 MB) > > This indicates that the results of each task are about 2705/1024 = 2.6MB > each. Is that normal? I don't know exactly what the result of each task > would be, but 2.6 MB for each seems too high. Can anyone offer an > explanation as to what the normal size should be if this is too high, or > ways to reduce this? > > Thanks. > > > > -- > View this message in context: http://apache-spark-user-list. > 1001560.n3.nabble.com/Serialized-task-result-size-exceeded-tp21449.html > Sent from the Apache Spark User List mailing list archive at Nabble.com. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: user-unsubscr...@spark.apache.org > For additional commands, e-mail: user-h...@spark.apache.org > >
