Chapter 9. Deploying Drill in Production
The two most common usage patterns for Drill are a single (often embedded) instance used to learn Drill, and a fully distributed, multinode setup used in production. This chapter explains the issues to consider when moving from personal usage to a production cluster. This chapter does not explain the basics of setting up a distributed Hadoop or Amazon Web Services (AWS) cluster; we assume that you already have that knowledge.
This section walks you through the steps required to get the Drill software on your nodes. Later sections explain how to configure Drill.
You have several options for how to install Drill on your servers:
- Vendor-provided installer
If you are a MapR customer, the easiest solution is to use the MapR installer. Note that the MapR configuration does not use the site directory (see “Creating a Site Directory”); if you upgrade Drill, be sure to follow the manual steps in the MapR documentation to save your configuration and JAR files before upgrading.
- Casual installation
If you are trying Drill on a server but are not quite ready for production use, you can install Drill much as you would on your laptop: just download Drill into your home directory and then follow the steps listed in Chapter 2.
- Production installation
This is a more structured way to install Drill onto your system; see “Production Installation”.
Regardless of how you do the installation, there are three common deployment patterns, ...