Hadoop speculative execution

You can found a good description about speculative execution at Wikipedia. Here are the two properties to configure the use of this feature:

mapred.map.tasks.speculative.execution
mapred.reduce.tasks.speculative.execution

Or if you are using Hadoop 2.x:

mapreduce.map.speculative
mapreduce.reduce.speculative

Most time it is useful but in some scenarios disabling it will make a big difference.

source

Related posts

Apr 22, '20

Running Apache Spark and S3 locally

3 min read
Apr 21, '20

Install Apache Spark on macOS

2 min read