In this chapter, we introduced the performance tuning process cycle and learned about Hadoop counters. We covered the TeraSort benchmark with its TeraGen module and learned to generate a performance baseline that will be used as a reference when tuning the Hadoop cluster. We also learned the approach to tune a Hadoop cluster illustrated by a three-node Hadoop cluster example and suggested tuning some setting parameters to improve the cluster's performance.

Then we moved ahead and discussed resource bottlenecks and what component is involved at each MapReduce stage or may potentially be the source of a bottleneck. For each component (CPU, RAM, storage, and network bandwidth), we learned how to identify system bottlenecks with suggestions to try to eliminate them.

In the next chapter, we will learn how to identify Hadoop cluster resource weaknesses and how to configure a Hadoop cluster correctly. Keep reading!