This is the documentation for CDH 4.7.0.
Documentation for other versions is available at Cloudera Documentation.

Usage Notes

Using the JobTracker Web UI

To use the JobTracker Web UI, use the HTTP address of either JobTracker (that is, the value of mapred.job.tracker.http.address.<name>.<id> for either the active or the standby JobTracker). Note the following:

  • If you use the URL of the standby JobTracker, you will be redirected to the active JobTracker.
  • If you use the URL of a JobTracker that is down, you will not be redirected - you will simply get a "Not Found" error from your browser.

Turning off Job Recovery

After a failover, the newly active JobTracker by default restarts all jobs that were running when the failover occurred. For Sqoop and HBase jobs, this is undesirable because they are not idempotent (that is, they do not behave the same way on repeated execution). For these jobs you should consider setting mapred.job.restart.recover to false in the job configuration (JobConf).