Total size of serialized results of 16 tasks (1048.5 MB) is bigger than spark.driver.maxResultSize (1024.0 MB)

Question

I get the following error when I add --conf spark.driver.maxResultSize=2050 to my spark-submit command.

17/12/27 18:33:19 ERROR TransportResponseHandler: Still have 1 requests outstanding when connection from /XXX.XX.XXX.XX:36245 is closed
17/12/27 18:33:19 WARN Executor: Issue communicating with driver in heartbeater
org.apache.spark.SparkException: Exception thrown in awaitResult:
        at org.apache.spark.util.ThreadUtils$.awaitResult(ThreadUtils.scala:205)
        at org.apache.spark.rpc.RpcTimeout.awaitResult(RpcTimeout.scala:75)
        at org.apache.spark.rpc.RpcEndpointRef.askSync(RpcEndpointRef.scala:92)
        at org.apache.spark.executor.Executor.org$apache$spark$executor$Executor$$reportHeartBeat(Executor.scala:726)
        at org.apache.spark.executor.Executor$$anon$2$$anonfun$run$1.apply$mcV$sp(Executor.scala:755)
        at org.apache.spark.executor.Executor$$anon$2$$anonfun$run$1.apply(Executor.scala:755)
        at org.apache.spark.executor.Executor$$anon$2$$anonfun$run$1.apply(Executor.scala:755)
        at org.apache.spark.util.Utils$.logUncaughtExceptions(Utils.scala:1954)
        at org.apache.spark.executor.Executor$$anon$2.run(Executor.scala:755)
        at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
        at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
        at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
        at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)
Caused by: java.io.IOException: Connection from /XXX.XX.XXX.XX:36245 closed
        at org.apache.spark.network.client.TransportResponseHandler.channelInactive(TransportResponseHandler.java:146)

The reason of adding this configuration was the error:

py4j.protocol.Py4JJavaError: An error occurred while calling o171.collectToPython.
: org.apache.spark.SparkException: Job aborted due to stage failure: Total size of serialized results of 16 tasks (1048.5 MB) is bigger than spark.driver.maxResultSize (1024.0 MB)

Therefore, I increased maxResultSize to 2.5 Gb, but the Spark job fails anyway (the error shown above). How to solve this issue?

Answer 1

Cause: caused by actions like RDD's collect() that send big chunk of data to the driver

Solution: set by SparkConf: conf.set("spark.driver.maxResultSize", "4g") OR set by spark-defaults.conf: spark.driver.maxResultSize 4g OR set when calling spark-submit: --conf spark.driver.maxResultSize=4g