请问有没有人分析过spark的输出日志？

2017-04-17 12:49:22  [ DataStreamer for file /historyserverforSpark/app-20170417124915-0005.inprogress block BP-1917352475-192.168.0.118-1490296260259:blk_1073768161_27337:12652 ] - [ DEBUG ]  DataStreamer block BP-1917352475-192.168.0.118-1490296260259:blk_1073768161_27337 sending packet packet seqno:2 offsetInBlock:21504 lastPacketInBlock:false lastByteOffsetInBlock: 26418
2017-04-17 12:49:22  [ ResponseProcessor for block BP-1917352475-192.168.0.118-1490296260259:blk_1073768161_27337:12657 ] - [ DEBUG ]  DFSClient seqno: 2 status: SUCCESS downstreamAckTimeNanos: 0
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12880 ] - [ INFO ]  Block broadcast_1 stored as values in memory (estimated size 4.1 KB, free 236.9 KB)
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12881 ] - [ DEBUG ]  Put block broadcast_1 locally took  10 ms
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12881 ] - [ DEBUG ]  Putting block broadcast_1 without replication took  10 ms
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12899 ] - [ INFO ]  Block broadcast_1_piece0 stored as bytes in memory (estimated size 2.3 KB, free 239.2 KB)
2017-04-17 12:49:22  [ dispatcher-event-loop-0:12900 ] - [ INFO ]  Added broadcast_1_piece0 in memory on 192.168.0.118:42100 (size: 2.3 KB, free: 511.1 MB)
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12905 ] - [ DEBUG ]  Updated info of block broadcast_1_piece0
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12905 ] - [ DEBUG ]  Told master about block broadcast_1_piece0
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12905 ] - [ DEBUG ]  Put block broadcast_1_piece0 locally took  6 ms
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12906 ] - [ DEBUG ]  Putting block broadcast_1_piece0 without replication took  6 ms
2017-04-17 12:49:22  [ dag-scheduler-event-loop:12906 ] - [ INFO ]  Created broadcast 1 from broadcast at DAGScheduler.scala:1006
2017-04-17 12:49:23  [ dag-scheduler-event-loop:12969 ] - [ INFO ]  Submitting 1600 missing tasks from ShuffleMapStage 0 (MapPartitionsRDD[3] at map at WordCount.scala:44)

请问在12：49：23提交1600个任务到集群之前为什么会有Block broadcast_1 stored as values in memory ，这个broadcast_1是做什么用的？

我觉像广播变量，将它放到内存了。

muyannian 发表于 2017-4-17 16:13
我觉像广播变量，将它放到内存了。

就是在集群上提交wordcount作业