CDH安装方法

CDH 5.3.1

Cloudera’s 100% Open Source Hadoop Platform

CDH is Cloudera's open source software distribution and consists of Apache Hadoop and additional key open source projects to ensure you get the most out of Hadoop and your data.

It is the only Hadoop solution to offer unified querying options (including batch processing, interactive SQL, text search, and machine learning) and necessary enterprise security features (such as role-based access controls).

Version: 

Please Select Version Please note: CDH requires manual installation from the command line.
For a faster, automated installation download Cloudera Manager.

• Installing CDH

• What's New

• System Requirements

• Documentation

Installing the Latest CDH 5 Release

  Important:

• If you use Cloudera Manager, do not use these command-line instructions.

• This information applies specifically to CDH 5.3.x. If you use an earlier version of CDH, see the documentation for that version located at Cloudera Documentation.

 

Ways To Install CDH 5

You can install CDH 5 in any of the following ways:

• Install Cloudera Manager, CDH, and managed services in a Cloudera Manager Deployment.  Note: Cloudera recommends that you use this automated method if possible.

• Or use one of the manual methods described below:

  If you use one of these manual methods rather than Cloudera Manager, the first (downloading and installing the "1-click Install" package) is recommended in most cases because it is simpler than building or adding a repository.

• Download and install the CDH 5 1-click Install" package; OR

• Add the CDH 5 repository; OR

• Build your own CDH 5 repository

• Install from a CDH 5 tarball — see, the next topic, "How Packaging Affects CDH 5 Deployment".

How Packaging Affects CDH 5 Deployment

Installing from Packages

• To install and deploy YARN, follow the directions on this page and proceed with Deploying MapReduce v2 (YARN) on a Cluster.

• To install and deploy MRv1, follow the directions on this page and then proceed with Deploying MapReduce v1 (MRv1) on a Cluster.

Installing from a Tarball

  Note: The instructions in this Installation Guide are tailored for a package installation, as described in the sections that follow, and do not cover installation or deployment from tarballs.

• If you install CDH 5 from a tarball, you will install YARN.

• In CDH 5, there is no separate tarball for MRv1. Instead, the MRv1 binaries, examples, etc., are delivered in the Hadoop tarball itself. The scripts for running MRv1 are in the bin-mapreduce1 directory in the tarball, and the MRv1 examples are in the examples-mapreduce1 directory.

Before You Begin Installing CDH 5 Manually

• The instructions on this page are for new installations. If you need to upgrade from an earlier release, see Upgrading from CDH 4 to CDH 5.

• For a list of supported operating systems, see CDH 5 Requirements and Supported Versions.

• These instructions assume that the sudo command is configured on the hosts where you will be doing the installation. If this is not the case, you will need the root user (superuser) to configure it.

  Note:

If you are migrating from MapReduce v1 (MRv1) to MapReduce v2 (MRv2, YARN), see Migrating from MapReduce 1 (MRv1) to MapReduce 2 (MRv2, YARN)for important information and instructions.

  Important:  Running Services

When starting, stopping and restarting CDH components, always use the service (8) command rather than running scripts in /etc/init.d directly. This is important because service sets the current working directory to / and removes most environment variables (passing only LANG and TERM), to create a predictable environment for the service. If you run the scripts in /etc/init.d, locally-set environment variables could produce unpredictable results. If you install CDH from RPMs, service will be installed as part of the Linux Standard Base (LSB).

  Important:  Java Development Kit:

• if you have not already done so, install the Oracle Java Development Kit (JDK); see Java Development Kit Installation.

High Availability

In CDH 5 you can configure high availability both for the NameNode and the JobTracker or Resource Manager.

• For more information and instructions on setting up a new HA configuration, see High Availability.  Important:

  If you decide to configure HA for the NameNode, do not install hadoop-hdfs-secondarynamenode. After completing the HDFS HA software configuration, follow the installation instructions under Deploying HDFS High Availability.

• To upgrade an existing configuration, follow the instructions under Upgrading to CDH 5.

Steps to Install CDH 5 Manually

Step 1: Add or Build the CDH 5 Repository or Download the "1-click Install" package.

• If you are installing CDH 5 on a Red Hat system, you can download Cloudera packages using yum or your web browser.

• If you are installing CDH 5 on a SLES system, you can download the Cloudera packages using zypper or YaST or your web browser.

• If you are installing CDH 5 on an Ubuntu or Debian system, you can download the Cloudera packages using apt or your web browser.

On Red Hat-compatible Systems

Use one of the following methods to add or build the CDH 5 repository or download the package on Red Hat-compatible systems.   Note:

Use only one of the three methods.

• Download and install the CDH 5 "1-click Install" package OR

• Add the CDH 5 repository OR

• Build a Yum Repository

Do this on all the systems in the cluster.

To download and install the CDH 5 "1-click Install" package:

1. Click the entry in the table below that matches your Red Hat or CentOS system, choose Save File, and save the file to a directory to which you have write access (it can be your home directory).

   OS Version	Click this Link
   Red Hat/CentOS/Oracle 5	Red Hat/CentOS/Oracle 5 link
   Red Hat/CentOS/Oracle 6	Red Hat/CentOS/Oracle 6 link

2. Install the RPM. For Red Hat/CentOS/Oracle 5:

   $ sudo yum --nogpgcheck localinstall cloudera-cdh-5-0.x86_64.rpm

   For Red Hat/CentOS/Oracle 6 (64-bit):

   $ sudo yum --nogpgcheck localinstall cloudera-cdh-5-0.x86_64.rpm

Now continue with Step 1a: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

  Note:  Make sure your repositories are up to dateBefore proceeding, make sure the repositories on each system are up to date:

sudo yum clean all

This ensures that the system repositories contain the latest software (it does not actually install anything).

OR: To add the CDH 5 repository:

Click the entry in the table below that matches your RHEL or CentOS system, navigate to the repo file for your system and save it in the /etc/yum.repos.d/directory.

For OS Version	Click this Link
RHEL/CentOS/Oracle 5	RHEL/CentOS/Oracle 5 link
RHEL/CentOS/Oracle 6 (64-bit)	RHEL/CentOS/Oracle 6 link

Now continue with Step 2: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

  Note:  Make sure your repositories are up to dateBefore proceeding, make sure the repositories on each system are up to date:

sudo yum clean all

This ensures that the system repositories contain the latest software (it does not actually install anything).

OR: To build a Yum repository:

If you want to create your own yum repository, download the appropriate repo file, create the repo, distribute the repo file and set up a web server, as described under Creating a Local Yum Repository.

Now continue with Step 2: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

  Note:  Make sure your repositories are up to dateBefore proceeding, make sure the repositories on each system are up to date:

sudo yum clean all

This ensures that the system repositories contain the latest software (it does not actually install anything).

On SLES Systems

Use one of the following methods to download the CDH 5 repository or package on SLES systems.   Note:

Use only one of the three methods.

• Download and install the CDH 5 "1-click Install" PackageOR

• Add the CDH 5 repositoryOR

• Build a SLES Repository

To download and install the CDH 5 "1-click Install" package:

1. Download the CDH 5 "1-click Install" package.

   Click this link, choose Save File, and save it to a directory to which you have write access (for example, your home directory).

2. Install the RPM:

   $ sudo rpm -i cloudera-cdh-5-0.x86_64.rpm

3. Update your system package index by running:

   $ sudo zypper refresh

Now continue with Step 2: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

OR: To add the CDH 5 repository:

1. Run the following command:

   $ sudo zypper addrepo -f http://archive.cloudera.com/cdh5/sles/11/x86_64/cdh/cloudera-cdh5.repo

2. Update your system package index by running:

   $ sudo zypper refresh

Now continue with Step 2: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

  Note:  Make sure your repositories are up to dateBefore proceeding to the next step, make sure the repositories on each system are up to date:

sudo zypper clean --all

This ensures that the system repositories contain the latest software (it does not actually install anything).

OR: To build a SLES repository:

If you want to create your own SLES repository, create a mirror of the CDH SLES directory by following these instructions that explain how to create a SLESrepository from the mirror.

Now continue with Step 2: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

  Note:  Make sure your repositories are up to dateBefore proceeding to the next step, make sure the repositories on each system are up to date:

sudo zypper clean --all

This ensures that the system repositories contain the latest software (it does not actually install anything).

On Ubuntu or Debian Systems

Use one of the following methods to download the CDH 5 repository or package.

• Download and install the CDH 5 "1-click Install" Package OR

• Add the CDH 5 repository OR

• Build a Debian Repository

  Note:

• Use only one of the three methods.

• There is an extra step if you are adding a repository on Ubuntu Trusty, as described below.

• Unless you are adding a repository on Ubuntu Trusty, don't forget to run apt-get update after downloading, adding, or building the repository.

To download and install the CDH 5 "1-click Install" package:

1. Download the CDH 5 "1-click Install" package:

   OS Version	Click this Link
   Wheezy	Wheezy link
   Precise	Precise link
   Trusty	Trusty link

2. Install the package by doing one of the following:

• Choose Open with in the download window to use the package manager.

• Choose Save File, save the package to a directory to which you have write access (for example, your home directory), and install it from the command line. For example:

  sudo dpkg -i cdh5-repository_1.0_all.deb

  Note:  Make sure your repositories are up to dateBefore proceeding to the next step, make sure the repositories on each system are up to date:

sudo apt-get update

This ensures that the system repositories contain the latest software (it does not actually install anything).

Now continue with Step 2: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

OR: To add the CDH 5 repository:

• Download the appropriate cloudera.list file by issuing one of the following commands. You can use another HTTP client if wget is not available, but the syntax may be different.  Important: Ubuntu 14.04 (Trusty)

  If you are running Ubuntu Trusty, you need to perform an additional step after adding the repository. See "Additional Step for Trusty" below.

  OS Version	Command
  Debian Wheezy	$ sudo wget 'http://archive.cloudera.com/cdh5/ubuntu/wheezy/amd64/cdh/cloudera.list' \     -O /etc/apt/sources.list.d/cloudera.list
  Ubuntu Precise	$ sudo wget 'http://archive.cloudera.com/cdh5/ubuntu/wheezy/amd64/cdh/cloudera.list' \     -O /etc/apt/sources.list.d/cloudera.list
  Ubuntu Lucid	$ sudo wget 'http://archive.cloudera.com/cdh5/ubuntu/lucid/amd64/cdh/cloudera.list' \     -O /etc/apt/sources.list.d/cloudera.list
  Ubuntu Trusty	$ sudo wget 'http://archive.cloudera.com/cdh5/ubuntu/trusty/amd64/cdh/cloudera.list' \     -O /etc/apt/sources.list.d/cloudera.list

  Note:  Make sure your repositories are up to dateUnless you are adding a repository on Ubuntu Trusty, make sure the repositories on each system are up to date before proceeding to the next step:

sudo apt-get update

This ensures that the system repositories contain the latest software (it does not actually install anything).

Additional step for Trusty

This step ensures that you get the right ZooKeeper package for the current CDH release. You need to prioritize the Cloudera repository you have just added, such that you install the CDH version of ZooKeeper rather than the version that is bundled with Ubuntu Trusty.

To do this, create a file at  /etc/apt/preferences.d/cloudera.pref with the following contents:

Package: * Pin: release o=Cloudera, l=Cloudera Pin-Priority: 501

  Note:

You do not need to run apt-get update after creating this file.

Now continue with Step 1a: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

OR: To build a Debian repository:

If you want to create your own apt repository, create a mirror of the CDH Debian directory and then create an apt repository from the mirror.

  Note:  Make sure your repositories are up to dateBefore proceeding to the next step, make sure the repositories on each system are up to date:

sudo apt-get update

This ensures that the system repositories contain the latest software (it does not actually install anything).

Now continue with Step 1a: Optionally Add a Repository Key, and then choose Step 3: Install CDH 5 with YARN, or Step 4: Install CDH 5 with MRv1; or do both steps if you want to install both implementations.

Step 2: Optionally Add a Repository Key

Before installing YARN or MRv1: (Optionally) add a repository key on each system in the cluster. Add the Cloudera Public GPG Key to your repository by executing one of the following commands:

• For Red Hat/CentOS/Oracle 5 systems:

  $ sudo rpm --import http://archive.cloudera.com/cdh5/redhat/5/x86_64/cdh/RPM-GPG-KEY-cloudera

• For Red Hat/CentOS/Oracle 6 systems:

  $ sudo rpm --import http://archive.cloudera.com/cdh5/redhat/6/x86_64/cdh/RPM-GPG-KEY-cloudera

• For all SLES systems:

  $ sudo rpm --import http://archive.cloudera.com/cdh5/sles/11/x86_64/cdh/RPM-GPG-KEY-cloudera

• For Ubuntu or Debian systems:

  OS Version	Command
  Debian Wheezy	$ wget http://archive.cloudera.com/cdh5/debian/wheezy/amd64/cdh/archive.key -O archive.key $ sudo apt-key add archive.key
  Ubuntu Precise	$ wget http://archive.cloudera.com/cdh5/ubuntu/precise/amd64/cdh/archive.key -O archive.key $ sudo apt-key add archive.key
  Ubuntu Lucid	$ wget http://archive.cloudera.com/cdh5/ubuntu/lucid/amd64/cdh/archive.key -O archive.key $ sudo apt-key add archive.key
  Ubuntu Trusty	$ wget http://archive.cloudera.com/cdh5/ubuntu/trusty/amd64/cdh/archive.key -O archive.key $ sudo apt-key add archive.key

This key enables you to verify that you are downloading genuine packages.

Step 3: Install CDH 5 with YARN

  Note:

Skip this step if you intend to use only MRv1. Directions for installing MRv1 are in Step 4.

To install CDH 5 with YARN:

  Note:

If you decide to configure HA for the NameNode, do not install hadoop-hdfs-secondarynamenode. After completing the HA software configuration, follow the installation instructions under Deploying HDFS High Availability.

1. Install and deploy ZooKeeper.  Important:

   Cloudera recommends that you install (or update) and start a ZooKeeper cluster before proceeding. This is a requirement if you are deploying high availability (HA) for the NameNode.

   Follow instructions under ZooKeeper Installation.

2. Install each type of daemon package on the appropriate systems(s), as follows.

   Where to install	Install commands
   Resource Manager host (analogous to MRv1 JobTracker) running:
   Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-yarn-resourcemanager
   SLES	sudo zypper clean --all; sudo zypper install hadoop-yarn-resourcemanager
   Ubuntu or Debian	sudo apt-get update; sudo apt-get install hadoop-yarn-resourcemanager
   NameNode host running:
   Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-hdfs-namenode
   SLES	sudo zypper clean --all; sudo zypper install hadoop-hdfs-namenode
   Ubuntu or Debian	sudo apt-get install hadoop-hdfs-namenode
   Secondary NameNode host (if used) running:
   Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-hdfs-secondarynamenode
   SLES	sudo zypper clean --all; sudo zypper install hadoop-hdfs-secondarynamenode
   Ubuntu or Debian	sudo apt-get install hadoop-hdfs-secondarynamenode
   All cluster hosts except the Resource Manager running:
   Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-yarn-nodemanager hadoop-hdfs-datanode hadoop-mapreduce
   SLES	sudo zypper clean --all; sudo zypper install hadoop-yarn-nodemanager hadoop-hdfs-datanode hadoop-mapreduce
   Ubuntu or Debian	sudo apt-get install hadoop-yarn-nodemanager hadoop-hdfs-datanode hadoop-mapreduce
   One host in the cluster running:
   Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-mapreduce-historyserver hadoop-yarn-proxyserver
   SLES	sudo zypper clean --all; sudo zypper install hadoop-mapreduce-historyserver hadoop-yarn-proxyserver
   Ubuntu or Debian	sudo apt-get install hadoop-mapreduce-historyserver hadoop-yarn-proxyserver
   All client hosts running:
   Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-client
   SLES	sudo zypper clean --all; sudo zypper install hadoop-client
   Ubuntu or Debian	sudo apt-get install hadoop-client

  Note:

The hadoop-yarn and hadoop-hdfs packages are installed on each system automatically as dependencies of the other packages.

Step 4: Install CDH 5 with MRv1

  Note:

If you are also installing YARN, you can skip any packages you have already installed in Step 3: Install CDH 5 with YARN.

Skip this step and go to Step 3: Install CDH 5 with YARN if you intend to use only YARN.

  Important: Before proceeding, you need to decide:

• Whether to configure High Availability (HA) for the NameNode and/or JobTracker; see the High Availability for more information and instructions.

• Where to deploy the NameNode, Secondary NameNode, and JobTracker daemons. As a general rule:

• The NameNode and JobTracker run on the same "master" host unless the cluster is large (more than a few tens of nodes), and the master host (or hosts) should not run the Secondary NameNode (if used), DataNode or TaskTracker services.

• In a large cluster, it is especially important that the Secondary NameNode (if used) runs on a separate machine from the NameNode.

• Each node in the cluster except the master host(s) should run the DataNode and TaskTracker services.

If you decide to configure HA for the NameNode, do not install hadoop-hdfs-secondarynamenode. After completing the HA software configuration, follow the installation instructions under Deploying HDFS High Availability.

First, install and deploy ZooKeeper.   Important:

Cloudera recommends that you install (or update) and start a ZooKeeper cluster before proceeding. This is a requirement if you are deploying high availability (HA) for the NameNode or JobTracker.

Follow instructions under ZooKeeper Installation. Make sure you create the myid file in the data directory, as instructed, if you are starting a ZooKeeper ensemble after a fresh install.

Next, install packages.

Install each type of daemon package on the appropriate systems(s), as follows.   Note:

On Ubuntu systems, Ubuntu may try to start the service immediately after you install it. This should fail harmlessly, but if you want to prevent it, there is advice here.

Where to install	Install commands
JobTracker host running:
Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-0.20-mapreduce-jobtracker
SLES	sudo zypper clean --all; sudo zypper install hadoop-0.20-mapreduce-jobtracker
Ubuntu or Debian	sudo apt-get update; sudo apt-get install hadoop-0.20-mapreduce-jobtracker
NameNode host running:
Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-hdfs-namenode
SLES	sudo zypper clean --all; sudo zypper install hadoop-hdfs-namenode
Ubuntu or Debian	sudo apt-get install hadoop-hdfs-namenode
Secondary NameNode host (if used) running:
Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-hdfs-secondarynamenode
SLES	sudo zypper clean --all; sudo zypper install hadoop-hdfs-secondarynamenode
Ubuntu or Debian	sudo apt-get install hadoop-hdfs-secondarynamenode
All cluster hosts except the JobTracker, NameNode, and Secondary (or Standby) NameNode hosts running:
Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-0.20-mapreduce-tasktracker hadoop-hdfs-datanode
SLES	sudo zypper clean --all; sudo zypper install hadoop-0.20-mapreduce-tasktracker hadoop-hdfs-datanode
Ubuntu or Debian	sudo apt-get install hadoop-0.20-mapreduce-tasktracker hadoop-hdfs-datanode
All client hosts running:
Red Hat/CentOS compatible	sudo yum clean all; sudo yum install hadoop-client
SLES	sudo zypper clean --all; sudo zypper install hadoop-client
Ubuntu or Debian	sudo apt-get install hadoop-client

Step 5: (Optional) Install LZO

If you decide to install LZO ( Lempel–Ziv–Oberhumer compression), proceed as follows. For information about choosing a compression format, see  Choosing a Data Compression Format   Note:If you are upgrading to a new version of LZO, rather than installing it for the first time, you must first remove the old version; for example, on a RHEL system:

yum remove hadoop-lzo

1. Add the repository on each host in the cluster. Follow the instructions for your OS version:

   For OS Version	Do this
   Red Hat/CentOS/Oracle 5	Navigate to this link and save the file in the /etc/yum.repos.d/ directory.
   Red Hat/CentOS 6	Navigate to this link and save the file in the /etc/yum.repos.d/ directory.
   SLES
   Ubuntu or Debian	Navigate to this link and save the file as /etc/apt/sources.list.d/gplextras.list.  Important: Make sure you do not let the file name default to cloudera.list, as that will overwrite your existing cloudera.list.

1. Run the following command:

    $ sudo zypper addrepo -f  http://archive.cloudera.com/gplextras5/sles/11/x86_64/gplextras/ cloudera-gplextras5.repo

2. Update your system package index by running:

    $ sudo zypper refresh

2. Install the package on each host as follows:

   For OS version	Install commands
   Red Hat/CentOS compatible	sudo yum install hadoop-lzo
   SLES	sudo zypper install hadoop-lzo
   Ubuntu or Debian	sudo apt-get install hadoop-lzo

3. Continue with installing and deploying CDH. As part of the deployment, you will need to do some additional configuration for LZO, as shown underConfiguring LZO .  Important: Make sure you do this configuration after you have copied the default configuration files to a custom location and set alternatives to point to it.

Step 6: Deploy CDH and Install Components

Now proceed with:

• deploying CDH 5

• installing components.