hive sql设置环境为：
设置spark.shuffle.memoryFraction = 0.1
设置spark.storage.memoryFraction = 0.2“
设置spark.driver.memory = 8G“
设置spark.executor.memory = 32G” 设置hive.exec.reducers.bytes.per.reducer = 640000000“;
“设置spark.sql.windowExec.buffer.spill.threshold = 1500000”;
“设置spark.yarn.executor.memoryoverhead = 4096m”;
“设置spark.executor.cores = 1”;
“设置spark.rdd.compress = true”;
“设置hive.optimize.sort.dynamic.partition = true”;
“设置spark.executor.instances = 6”;
“设置spark.yarn.am.memory = 8G”;
“设置yarn.app.mapreduce.am.resource.mb = 8G”;
“设置spark.executor.heartbeatInterval = 240000”;
将运行的sqlString插入到tmptable中的historyTable中：
“插入表” + historyTable +“ partition（found_time）”中选择* *从“ + temporaryTable +”中，found_time不为null时，通过rand（）分发“
我的机器是： 3节点-双CPU 128内存
脚本在纱线上火花运行，程序可以稳定运行约2分钟，下一条错误消息仅见：
错误执行程序。CoarseGrainedExecutor后端：接收到的信号项