When i increase my input data size, the executor will be failed and lost. see below:
14/03/11 20:44:18 INFO AppClient$ClientActor: Executor updated: app-20140311204343-0008/8 is now FAILED (Command exited with code 134) 14/03/11 20:44:18 INFO SparkDeploySchedulerBackend: Executor app-20140311204343-0008/8 removed: Command exited with code 134 14/03/11 20:44:18 INFO SparkDeploySchedulerBackend: Executor 8 disconnected, so removing it 14/03/11 20:44:18 ERROR TaskSchedulerImpl: Lost executor 8 on Salve10.Hadoop: Unknown executor exit code (134) (died from signal 6?) 14/03/11 20:44:18 INFO TaskSetManager: Re-queueing tasks for 8 from TaskSet 1.0 14/03/11 20:44:18 WARN TaskSetManager: Lost TID 8 (task 1.0:22) 14/03/11 20:44:18 WARN TaskSetManager: Lost TID 49 (task 1.0:49) 14/03/11 20:44:18 WARN TaskSetManager: Lost TID 31 (task 1.0:40) 14/03/11 20:44:18 WARN TaskSetManager: Lost TID 1 (task 1.0:10) 14/03/11 20:44:18 WARN TaskSetManager: Lost TID 15 (task 1.0:48) 14/03/11 20:44:18 INFO AppClient$ClientActor: Executor added: app-20140311204343-0008/11 on worker-20140311172513-Salve10.Hadoop-56435 (Salve10.Hadoop:56435) with 8 cores 14/03/11 20:44:18 INFO SparkDeploySchedulerBackend: Granted executor ID app-20140311204343-0008/11 on hostPort Salve10.Hadoop:56435 with 8 cores, 4.0 GB RAM 14/03/11 20:44:18 INFO AppClient$ClientActor: Executor updated: app-20140311204343-0008/11 is now RUNNING It seems that too many tasks are running in a container and the total task memory exceed the "spark.executor.memory" so how to set the task number in a executor? thank you