我在保存大数据到hdfs时出现OOMEvalaccumulableCollection=sc.accumulableCollection(ArrayBuffer[String]())valrdd=textfile.filter(row=>{if(row.endsWith(",")){accumulableCollection+=rowfalse}elseif(row.length{varvalid=truefor((k,v)我在spark-submit中使用这个:--num-executors2--driver-memory1G--executor-memory1G--executor