Reputation: 5367
I'm using Hadoop Map/Reduce Administration UI coming with standard Apache distro. With the time the list of executed jobs is growing too big. How I can remove old jobs from the history list?
Upvotes: 1
Views: 3124
Reputation: 30089
Another suggestion (more extreme) is just to restart your jobtracker.
Upvotes: 0
Reputation:
Hadoop job execution has life cycle, which is explained
here
Job status flow
Running -> Completed -> Retired
There are important configurations using which we can control the job retention.
mapred.jobtracker.completeuserjobs.maximum: Number of jobs to keep in memory after Running state. Try configure to less value...
mapred.jobtracker.retirejob.interval: Default interval value is 24 * 60 * 60 * 1000 (1 Day). That means completed jobs will be retired after one day by default.
Try configuring property "mapred.jobtracker.retirejob.interval" to 1 hour or so.
Upvotes: 1