cron任务：每隔1分钟启动脚本，检查进程是否运行。crontab -e
*/1 * * * * /data/spark/test.sh检查进程，如果进程挂掉，重新启动Spark任务：
#!/bin/sh
is_Engine_exist=$(ps aux | grep LbsStreamingEngineTJ | grep -v grep | wc -l)
if [ $is_Engine_exist = 0 ];then
echo ‘Process Engine is down‘
echo ‘Bring Engine up‘
strDate=`date +%Y%m%d%H%M%S`
strStart="start Engine ${strDate}"
echo "${strStart}" >> /data1/log.txt
nohup /data1/spark-1.6.0/bin/spark-submit --master spark://localhost:7077 --name LbsStreamingEngineTJ --class com.datafactory.streaming.LbsStreamingEngineTJ --executor-memory 512m --total-executor-cores 2 /data1/work/datafactory-0.1.0-SNAPSHOT1023.jar &
echo ‘Bring Engine finished ‘
else
strDate=`date +%Y%m%d%H%M%S`
strRun="running ${strDate}"
echo "${strRun}" >> /data1/log.txt
fi
SparkStreaming任务保持运行，定时任务监控进程，保证不挂掉
原文：https://www.cnblogs.com/Bread-Wang/p/9957294.html