by Alex McLintock and Alan Duval of Alephant.co.uk, Sept 2017
This is the introduction to a series of articles about installing Cloudera's Distribution of Hadoop (CDH version 5.12) on CentOS 7.3 machines by first installing and using Cloudera Manager.
Cloudera produces a lot of documentation, but we find that it makes a lot of assumptions about your knowledge and can be confusing. Here we are providing a guide through one possible way of installing Cloudera Manager and CDH on CentOS machines. Each time we come to a decision point in the installation documentation we try to make that clear and explain the choice we have taken. We still link to the relevant Cloudera pages throughout so you can read up more if need be. Most of these web pages refer to the "latest" version of CDH - which is 5.12 at the time of writing - so the content may change when a new CDH is released.
We detail, in separate articles:
- Preparing to install Cloudera CDH - what you need to do first, including preparing an optional package repository
- Doing the Install: Cloudera CDH5.12 on CentOS7.3 - how to do the install once everything is prepared. Instructions for RedHat will be similar.
- Smoke Testing Installed CDH - very basic testing that what you have installed works
Each of these articles will be expanded over time with future articles. Do ask us if there are any aspects you are keen to read more about
CDH vs Cloudera Manager
Cloudera Manager is the tool which normally installs CDH and does administration of it.
Some time ago Cloudera Manager was seen as a part of CDH, but more recently it has been separated out. We try to follow this convention, but in some cases we may say CDH and include Cloudera Manager within that.
Despite being separate tools, Cloudera Manager and CDH both have the same version: 5.12. A new version of Cloudera Manager is released with each new version of CDH. It is possible to upgrade Cloudera Manager to the latest version without upgrading CDH, but not vice versa. According to CDH 5 and Cloudera Manager 5 Requirements and Supported Versions the Cloudera Manager minor version number (in this case .12) must always be equal to or higher than the version number for the CDH that it is managing.
Article Formatting Style
We hope that everything is clear but what follows is an explanation of our house style:
"Note" - will be used when we are making commentary about the preceding line(s) of text. Particularly where we feel that Cloudera's documentation is a little confusing and we are trying to clarify what we believe was intended.
"! -" will be used when we are discussing a point in the installation that requires an important decision between two or more options.
"? -" will be used where we encounter something we're not 100% sure about, and are addressing a question to you, the reader. Feel free to respond in the comments.
Code will be in a fixed width font - sometimes in blocks
Italics denote keywords that we need to highlight.
For URLs such as Cloudera Product Documentation, we will use the hyperlinked name of the document to aid readability. We will repeat the URL at the end of that block of text to help with printouts.
You should read Preparing to install Cloudera CDH before charging in and doing the install.