Deploying the SAS Embedded Process Stack on Hortonworks

Deploying the SAS Embedded Process Stack for the First Time
Deploying a New Version of the SAS Embedded Process Stack

Deploying the SAS Embedded Process Stack for the First Time

After you run the SAS Deployment Manager to create the SAS Embedded Process stack, you must deploy the stack on the cluster. Follow these steps:
Note: If the SAS Embedded Process stack already exists on your cluster, follow the instructions in Deploying a New Version of the SAS Embedded Process Stack.
  1. Start the Ambari server and log on.
  2. If the requiretty option was enabled when you deployed the SAS Embedded Process, you must restart the Ambari server at this time. Otherwise, skip to Step 3.
    1. Log on to the cluster.
      sudo - su
    2. Restart the Ambari server.
      ambari-server restart
    3. Start the Ambari server and log on.
  3. Click Actions and choose + Add Service.
    The Add Service Wizard page appears.
  4. Select Choose Services.
    The Choose Services panel appears.
  5. In the Choose Services panel, select SASEP SERVICE. Click Next.
    The Assign Slaves and Clients panel appears.
    Note: You should always select NAMENODE as one of the clients and NAMENODE should have these two client components installed: HDFS_CLIENT and HCAT_CLIENT.
    Note: If you have licensed and downloaded SAS Data Loader for Hadoop, some SAS Data Loader for Hadoop data quality components are silently deployed at the same time as the SAS Embedded Process for Hadoop. Other configuration is required as noted in step 11.
  6. In the Assign Slaves and Clients panel, select items under Client where you want the stack to be deployed.
    The Customize Services panel appears.
    The SASEP stack is listed under activated_version. The stack name is the one from Step 6 in Using the SAS Deployment Manager to Create the SAS Embedded Process Parcel or Stack. An example name is 9.43.s0.1.
  7. Do not change any settings on the Customize Services panel. Click Next.
    If your cluster is secured with Kerberos, the Configure Identities panel appears. Enter your Kerberos credentials in the admin_principal and admin_password text boxes.
    The Review panel appears.
  8. Review the information about the panel. If everything is correct, click Deploy.
    The Install, Start, and Test panel appears. When the SAS Embedded Process stack is installed on all nodes, click Next.
    The Summary panel appears.
  9. Click Complete. The SAS Embedded Process stack is now installed on all nodes of the cluster.
    You should now be able to see SASEP SERVICE on the Ambari dashboard.
  10. Verify that the ep-config.xml file exists in the /sas/ep/config directory.
  11. Review any additional configuration that might be needed depending on your Hadoop distribution.
  12. Validate the deployment of the SAS Embedded Process by running a program that uses the SAS Embedded Process and the MapReduce service. An example is a scoring program.
  13. If you have licensed and downloaded any of the following SAS software, additional configuration is required:
    • SAS Data Loader for Hadoop
      For more information, see Part 3, “Administrator’s Guide for SAS Data Loader for Hadoop”.
    • SAS High-Performance Analytics
      For more information, see SAS High-Performance Analytics Infrastructure: Installation and Configuration Guide.

Deploying a New Version of the SAS Embedded Process Stack

More than one SAS Embedded Process stack can be deployed on your cluster, but only one stack can be activated at one time. After you run the SAS Deployment Manager to create the SAS Embedded Process stack, follow these steps to deploy an additional SAS Embedded Process stack when one already exists on your cluster.
  1. Restart the Ambari server and log on to the Ambari manager.
  2. Select SASEP SERVICE.
    In the Services panel, a restart symbol appears next to SASEP SERVICE. The Configs tab indicates that a restart is required.
  3. Click Restart.
  4. Click Restart All.
    After the service is restarted, the previous version of the SAS Embedded Process still appears in the activated_version text box on the Configs tab. All deployed versions of the SAS Embedded Process stack should appear in the sasep_allversions text box.
  5. Refresh the browser.
    The new version of the SAS Embedded Process should now appear as the activated_version text box on the Configs tab.
If, at any time, you want to activate another version of the SAS Embedded Process stack, follow these steps:
  1. Enter the version number in the activated_version text box on the Configs tab.
  2. Click Save.
  3. Add a note describing your action (for example, “Changed from version 9.43.s01.1 to 9.43.s01.2”), and click Next.
  4. Click Restart.
  5. Click Restart All.
  6. Refresh Ambari.
    The new service is activated.
  7. Review any additional configuration that might be needed depending on your Hadoop distribution.
  8. Validate the deployment of the SAS Embedded Process by running a program that uses the SAS Embedded Process and the MapReduce service. An example is a scoring program.
  9. If you have licensed and downloaded any of the following SAS software, additional configuration is required:
    • SAS Data Loader for Hadoop
      For more information, see Part 3, “Administrator’s Guide for SAS Data Loader for Hadoop”.
    • SAS High-Performance Analytics
      For more information, see SAS High-Performance Analytics Infrastructure: Installation and Configuration Guide.
Copyright © SAS Institute Inc. All rights reserved.