Chapter 9. Deploying HBase
This chapter covers
By now, you’ve learned a lot about HBase as a system and how to use it. As you’ve read the chapters, we hope you’ve set up HBase in standalone mode and played around with client code. A single-node standalone HBase install is only meant for basic access, which you typically do either while learning how to use the system or while developing an application. It can’t handle any real workload or scale.
When planning a fully distributed HBase setup, you have to think about all the individual components: HBase Master, ZooKeeper, RegionServers, and HDFS DataNodes. Sometimes this list also includes the MapReduce framework. Each of these has different requirements in terms of hardware resources. This chapter will teach you in detail about the requirements for all the components and how you should choose hardware for a fully distributed HBase install. We’ll then talk about the different HBase distributions available and considerations you should take into account when choosing one over the other. We’ll also discuss deployment strategies and what you should consider when architecting your deployment system.