Thursday, September 26, 2013

Hadoop's speculative task execution

One problem with the Hadoop system is that by dividing the tasks across many nodes, it is possible for a few slow nodes to rate-limit the rest of the program.

Tasks may be slow for various reasons, including hardware degradation, or software mis-configuration, but the causes may be hard to detect since the tasks still complete successfully, albeit after a longer time than expected. Hadoop doesn’t try to diagnose and fix slow-running tasks; instead, it tries to detect when a task is running slower than expected and launches another, equivalent, task as a backup. This is termed speculative execution of tasks.

For example if one node has a slow disk controller, then it may be reading its input at only 10% the speed of all the other nodes. So when 99 map tasks are already complete, the system is still waiting for the final map task to check in, which takes much longer than all the other nodes.

By forcing tasks to run in isolation from one another, individual tasks do not know where their inputs come from. Tasks trust the Hadoop platform to just deliver the appropriate input. Therefore, the same input can be processed multiple times in parallel, to exploit differences in machine capabilities. As most of the tasks in a job are coming to a close, the Hadoop platform will schedule redundant copies of the remaining tasks across several nodes which do not have other work to perform. This process is known as speculative execution. When tasks complete, they announce this fact to the JobTracker. Whichever copy of a task finishes first becomes the definitive copy. If other copies were executing speculatively, Hadoop tells the TaskTrackers to abandon the tasks and discard their outputs. The Reducers then receive their inputs from whichever Mapper completed successfully, first.

Speculative execution is enabled by default. You can disable speculative execution for the mappers and reducers by setting the and mapred.reduce.tasks.speculative.execution JobConf options to false, respectively using old API, while with newer API you may consider changing and mapreduce.reduce.speculative.

If your jobs' mapper/redcuer takes very less time then it is better to disable the speculative execution. But in cases where your jobs' each mapper/reducer takes a considerable amount of time, you would be better off having speculative execution enabled as if after running for say 45 minutes, if a mapper/reducer fails due to loss of that particular node then another node has to restart it's execution from scratch.



  1. You have certainly explained that mapred is the process of examining big data to uncover hidden patterns, unknown correlations and other useful information that can be used to make better decisions..The big data analytics is the major part to be understood regarding Best Hadoop Training in Chennai program. Via your quality content i get to know about that in deep.Thanks for sharing this here.

  2. The content provided here is vital in increasing one's knowledge regarding hadoop, the way you have presented here is simply awesome. Thanks for sharing this. The uniqueness I see in your content made me to comment on this. Keep sharing article like this. Thanks :)

    Hadoop Training Chennai | Big Data Training in Chennai | hadoop course in Chennai

  3. Hi, Really your post was very informative. Today's internet era learn Hadoop Online Training will helps you to reach your goal.Selenium Online Training

  4. I simply wanted to write down a quick word to say thanks to you for those wonderful tips and hints you are showing on this site.

    big data training in chennai
    hadoop training in bangalore

  5. Really useful information about hadoop, i have to know information about hadoop online training institutes.

  6. I wish to show thanks to you just for bailing me out of this particular
    trouble.As a result of checking through the net and meeting
    techniques that were not productive, I thought my life was done.

    java training in chennai


Any feedback, good or bad is most welcome.


Email *

Message *