1.问题
最近发现SparkStreaming提交的job经常在半夜挂掉,于是写了个定时任务监控SparkStreaming的运行状态,保证其不挂掉
2.shell脚本
touch /opt/module/jobs/monitorlog.txt vim /opt/module/jobs/monitor.sh
#!/bin/bash #在linux中查找你所运行的spark任务中 任务名称为WBStreamingClusterDriver的任务有没有,如果有则返回值是1 job_status=$(yarn application -list| awk '{print $2}' | grep KeyBehaviorsQl | wc -l) if [ $job_status = 0 ];then echo $(date "+%Y-%m-%d %H:%M:%S") 'SparkStreamingTestis stop' >> /opt/module/jobs/monitorlog.txt nohup spark2-submit --master yarn --deploy-mode cluster --driver-memory 3g --num-executors 70 \ --executor-cores 2 --executor-memory 3g \ --class com.eighteen.sparkstreaming.SparkStreamingTest\ /opt/module/jobs/SparkStreaming-1.0-SNAPSHOT-jar-with-dependencies.jar echo $(date "+%Y-%m-%d %H:%M:%S") "restart SparkStreamingTest success !!!" >> /opt/module/jobs/monitorlog.txt else echo $(date "+%Y-%m-%d %H:%M:%S") 'SparkStreamingTest is running !!' >> /opt/module/jobs/monitorlog.txt fi
3.编写定时任务
*/3 * * * * /opt/module/jobs/monitor.sh