You are viewing a plain text version of this content. The canonical link for it is here.

Posted to mapreduce-dev@hadoop.apache.org by Yoonmin Nam <ro...@dgist.ac.kr> on 2013/12/14 09:03:46 UTC

How can I get the spill time of each task?

Hi.

 

I want to get exact time spent to spill a map output into disk.

Especially, the time spent to flush method in MapTask.java

 

It actually easy job when I use a pseudo-distributed mode, but

In real cluster, I tried to use custom counter in MapTask and ReduceTask, it
isn't work.

Also, stdout isn't working.

 

So, I want you ask about above issues.

 

Thanks!