Note: For more
information about the SAS Deployment Manager pages, click Help on
each page.
-
Start the SAS Deployment
Manager.
cd /SASHOME/SASDeploymentManager/9.4
./sasdm.sh
The Choose
Language page opens.
-
Select the language
in which you want to perform the configuration of your software.
Click OK.
The Select SAS Deployment Manager Task page
opens.
-
Under Hadoop Configuration,
select Deploy SAS In-Database Technologies for Hadoop.
Click Next to
continue. The Select Hadoop Distribution page
opens.
Note: If you
have licensed and downloaded SAS Data Loader for Hadoop, the SAS Data
Quality Accelerator component is silently deployed at the same time
as the SAS Embedded Process for Hadoop.
Note: If you
have licensed and downloaded SAS Contextual Analysis In-Database Scoring
for Hadoop, the SAS Contextual Analysis In-Database Scoring for Hadoop
component is silently deployed at the same time as the SAS Embedded
Process for Hadoop.
-
From the drop-down menu,
select the distribution of Hadoop that you are using.
Note: If your distribution is not
listed, exit the SAS Deployment Manager and contact SAS Technical
Support.
Click Next.
The Hadoop Cluster Manager Information page
opens.
-
Enter the host name
and port number for your Hadoop cluster.
For Cloudera, enter
the location where Cloudera Manager is running. For Hortonworks, IBM
BigInsights, or Pivotal HD, enter the location where the Ambari server
is running.
The port number is set
to the appropriate default after Cloudera, Hortonworks, IBM BigInsights,
or Pivotal HD, is selected.
Note: The host name must be a fully
qualified domain name. The port number must be valid, and the cluster
manager must be listening.
Click Next.
The Hadoop Cluster Manager Credentials page
opens.
-
Enter the Cloudera Manager
or Ambari administrator account name and password.
Note: Using the credentials of
the administrator account to query the Hadoop cluster and to find
the Hive node eliminates guesswork and removes the chance of a configuration
error. However, the account name does not have to be that of an administrator;
it can be a read-only user.
Click Next.
The UNIX
User Account with SSH for the Hadoop Cluster Manager Host
page opens.
-
Enter the root SSH account
that has access to the cluster manager or enter a non-root SSH account
if that account can execute sudo without entering a password.
Note: For Cloudera, the SSH account
must have Write permission to the /opt/cloudera
directory.
Otherwise, the deployment completes with errors.
Click Next.
The Specify the SAS Configuration and Deployment Directories page
opens.
-
Enter the location of
the SAS configuration and deployment directories:
-
Enter (or navigate to)
the location of the /standalone_installs
directory.
This directory was created when your SAS Software Depot was created
by the SAS Download Manager.
CAUTION:
After installation,
do not delete your SAS Software Depot standalone_installs directory
or any of its subdirectories.
If hot fixes are made
available for your software, they are moved to a subdirectory of the /standalone_installs/SAS_Core_Embedded_Process_Package_for_Hadoop/
directory.
The SAS Deployment Manager requires that both the initial installation
files and the hot fix file exist in a subdirectory of the original
SAS Software Depot /standalone_installs/SAS_Core_Embedded_Process_Package_for_Hadoop/
directory.
-
Enter (or navigate to)
a working directory on the local server where the package or stack
is placed. The working directory is removed when the deployment is
complete.
Click Next.
The Checking System page opens, and a check
for locked files and Write permissions is performed.
-
If any files are shown
in the text box after the system check, follow the instructions on
the Checking System page to fix any problems.
Click Next.
The Summary page opens.
-
Click Start to
begin the configuration.
Note: It takes time to complete
the configuration. If your cluster is secured with Kerberos, it could
take longer.
Note: The product
that appears on this page is the SAS product that is associated with
the in-database deployment package for Hadoop. This package includes
the SAS Embedded Process and possibly other components. Note that
a separate license might be required to use the SAS Embedded Process.
If the configuration
is successful, the page title changes to Deployment Complete and
a green check mark is displayed beside SAS In-Database Technologies
for Hadoop (64-bit).
Note: Part of the configuration
process runs SAS code to validate the environment. A green check mark
indicates that the SAS Deployment Manager was able to create the SAS
Embedded Process parcel or stack and then verify that the parcel or
stack was copied to the cluster manager node.
If warnings or errors
occur, fix the issues and restart the configuration.
-
Click Next to
close the SAS Deployment Manager.
A log file is written
to the %HOME/.SASAppData/SASDeploymentWizard
directory
on the client machine.
-
Continue the installation
process.