我遇到Apache Spark(Graphx)内存不足的问题。应用程序运行,但一段时间后关闭。我使用Spark 1.2.0。群集有足够的内存和多个内核。我没有使用GraphX的其他应用程序,运行没有问题。申请使用Pregel。
我以Hadoop YARN模式提交申请:
HADOOP_CONF_DIR = / etc / hadoop / conf spark-submit --class DPFile --deploy-mode cluster --master yarn --num-executors 4 --driver-memory 10g --executor-memory 6g --executor-核心8 - 文件log4j.properties spark_routing_2.10-1.0.jar road_cr_big2 1000
Spark配置:
var intvalue = (int)test[1];
8207
感谢您的回答。
日志:
val conf = new SparkConf(true)
.set("spark.eventLog.overwrite", "true")
.set("spark.driver.extraJavaOptions", "-Dlog4j.configuration=log4j.properties")
.set("spark.yarn.applicationMaster.waitTries", "60")
.set("yarn.log-aggregation-enable","true")
.set("spark.akka.frameSize", "500")
.set("spark.akka.askTimeout", "600")
.set("spark.core.connection.ack.wait.timeout", "600")
.set("spark.akka.timeout","1000")
.set("spark.akka.heartbeat.pauses","60000")
.set("spark.akka.failure-detector.threshold","3000.0")
.set("spark.akka.heartbeat.interval","10000")
.set("spark.ui.retainedStages","100")
.set("spark.ui.retainedJobs","100")
.set("spark.driver.maxResultSize","4G")