为什么减少卡在 16%？

Question

我有一个 map reduce 作业，我试图在一个相对较小的数据集上运行。我一直遇到一个问题，即减少工作一直卡在 16%。我的任务跟踪器的日志显示：

2012-03-21 17:09:23,829 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:26,865 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:32,902 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:38,938 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:41,973 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:48,010 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:51,045 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:09:57,086 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:03,120 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:06,154 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:12,198 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:18,234 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:21,271 INFO org.apache.hadoop.mapred.TaskTracker:> attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:27,310 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:33,342 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:36,374 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:42,403 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:48,435 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:51,462 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:10:57,495 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:11:03,523 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:11:06,545 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:11:12,578 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)
2012-03-21 17:11:18,607 INFO org.apache.hadoop.mapred.TaskTracker: attempt_201203211704_0001_r_000000_0 0.16666667% reduce > copy (1 of 2 at 0.16 MB/s)

Answer 1

我打开了一个 FileSystem 对象，但没有在访问文件的映射器中关闭它。在 map() 定义的末尾添加 fs.close() 解决了这个问题。