[英]EMR cluster hangs in Step state 'Running/Pending'
我正在使用自定义 jar 步骤通过 java SDK 启动 EMR 集群。 集群启动成功,但是当步骤处于挂起/运行 state 的引导后,集群卡住了。 我什至无法在机器上使用 ssh。
以下是我使用自定义 jar 步骤启动集群的代码 -
String dataTrasnferJar = s3://test/testApplication.jar;
if (dataTrasnferJar == null || dataTrasnferJar.isEmpty())
throw new InvalidS3ObjectException(
"EMR custom jar file path is null/empty. Please provide a valid jar file path");
HadoopJarStepConfig customJarConfig = new HadoopJarStepConfig().withJar(dataTrasnferJar);
StepConfig customJarStep = new StepConfig("Mongo_to_S3_Data_Transfer", customJarConfig)
.withActionOnFailure(ActionOnFailure.CONTINUE);
AmazonElasticMapReduce emr = AmazonElasticMapReduceClientBuilder.standard()
.withCredentials(awsCredentialsProvider)
.withRegion(region)
.build();
Application spark = new Application().withName("Spark");
String clusterName = "my-cluster-" + System.currentTimeMillis();
RunJobFlowRequest request = new RunJobFlowRequest()
.withName(clusterName)
.withReleaseLabel("emr-6.0.0")
.withApplications(spark)
.withVisibleToAllUsers(true)
.withSteps(customJarStep)
.withLogUri(loggingS3Bucket)
.withServiceRole("EMR_DefaultRole")
.withJobFlowRole("EMR_EC2_DefaultRole")
.withInstances(new JobFlowInstancesConfig()
.withEc2KeyName(key_pair)
.withInstanceCount(instanceCount)
.withEc2SubnetIds(subnetId)
.withAdditionalMasterSecurityGroups(securityGroup)
.withKeepJobFlowAliveWhenNoSteps(true)
.withMasterInstanceType(instanceType));
RunJobFlowResult result = emr.runJobFlow(request);
EMR emr-6.0.0 版本仍在开发中。 您可以对 emr-5.29.0 进行相同的尝试吗?
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.