Chapter 4. HDFS Details for Multimachine Clusters
As you learned in the previous chapter, the defaults provided for multimachine clusters will work well for very small clusters, but they are not suitable for large clusters (the clusters will fail in unexpected and difficult-to-understand ways). This chapter covers HDFS installation for multimachine clusters that are not very small, as well as HDFS tuning factors, recovery procedures, and troubleshooting tips. But first, let's look at some of the configuration trade-offs faced by IT departments.
There appears to be an ongoing conflict between the optimal machine and network configurations for Hadoop Core and the configurations required by IT departments.
IT departments are ...