I start a job on a Hadoop cluster using JobClient
, which gives me a handle to a RunningJob
. Is there a painless way to get the log output of just that particular job? Or do I have to write some code to dig through the logs directory and match the names against the job ID?
Alternately, is there a way to redirect the log output per individual job, or is all logging in Hadoop configu开发者_JAVA百科red globally?
You can have the job's specific output go to the console of your running job if you code output you want using System.out.println inside of your Java code otherwise you would have to grep for your JobID in the Hadoop logs.
You can also use counters inside of your job http://philippeadjiman.com/blog/2010/01/07/hadoop-tutorial-series-issue-3-counters-in-action/
精彩评论