Defining Multiple Application Servers

Overview of Adding an Additional SAS Application Server

When you deploy SAS, the SAS Deployment Wizard creates an initial SAS Application Server using the inputs that you supply. Later, you can re-run the SAS Deployment Wizard to create additional application servers on new machines. For more information, see Adding SAS Products in SAS Intelligence Platform: Installation and Configuration Guide.

Follow these guidelines when running the SAS Deployment Wizard to add a new SAS Application Server:

The machine on which you want to add the new logical server should have a network connection to your site's SAS Software Depot. If it does not, then you must find a way to get the depot to the machine on which you want to create the new SAS Application Server.

For more information, see Overview of SAS Software Depots in SAS Intelligence Platform: Installation and Configuration Guide.
Use your deployment plan that contains the SAS Application Server.

Note: There are standard deployment plans that contain a SAS Application Server machine. For more information, see About Deployment Plans in SAS Intelligence Platform: Installation and Configuration Guide.
If you are adding logical servers that require an object spawner (one of the workspace servers and the stored process server), then when you deploy your SAS Application Server, the wizard also deploys an object spawner to start the servers. If the machine on which you are adding the SAS Application Server already contains an object spawner, then the wizard updates the pre-existing spawner definition for you to include the new servers that you want your spawner to manage.
Specify a new server context (for example, SASApp2).
Obtain the SAS Metadata Server connection information for the metadata server that you are currently using. The wizard prompts you for this information.
Specify the correct machine name on which you are adding the SAS Application Server.
After the server is registered in metadata and running, make sure that Job Execution Service is configured to use the newly registered server. For more information, see Job Execution Services in SAS Intelligence Platform: Middle-Tier Administration Guide.

There are several reasons why you might want to create a second application server. A couple of these reasons are discussed in the following sections.

Scenario 1: Using SAS Data Integration Studio to Access Remote Data

Suppose that you are using SAS Data Integration Studio to process a large amount of data that resides on a machine different from the workspace server to which the application submits its code, as shown in the following figure.

Application Server and Data on Different Machines

One way to execute such a job efficiently is to define two application servers. One is the default application server for SAS Data Integration Studio and contains the workspace server to which the application submits its generated code. The other application server contains a SAS/CONNECT server (which is collocated with the data), and the library of data to be processed is assigned to this application server. See the following figure.

Create an Application Server That Is Collocated with the Data

If you choose for the transformations in the job to be executed on the remote host, then SAS Data Integration Studio generates the code that is necessary for the transformations to be executed by the SAS/CONNECT server. The overall job is submitted to the workspace server. However, the workspace server then submits the code for each transformation to the remote server.

The generation of this type of code is made possible by the definitions of the two application servers, one of which is the default application server for SAS Data Integration Studio and the other the application server to which the data library is assigned.

Note: For more information, see Setting Up Multi-Tier Environments in SAS Intelligence Platform: Desktop Application Adminstration Guide.

Scenario 2: Using Multiple Application Servers with SAS Web Report Studio

Suppose that you want to create an environment in which row-level security can be strictly enforced for a set of SAS Web Report Studio users. (For a detailed description of this feature, see Overview of BI Row-Level Permissions in SAS Guide to BI Row-Level Permissions.)

Part of the setup is to create a special workspace server for use by the report creators who need the secure environment. This special workspace server is a component of a new SAS Application Server. See the following figure.

Using a Second Application Server to Enable Row-Level Security

The original workspace server can service users of other applications and users of SAS Web Report Studio whose access to data does not need to be so closely controlled.

Create a SAS Application Server for a SAS Solution

SAS solutions such as SAS Demand-Driven Forecasting require unique SAS Application Servers in order to properly function.

If you are deploying your SAS solution on a system without any pre-existing SAS software, you can proceed normally making sure to name the SAS Application Server Context for your solution. For more information, see the SAS Intelligence Platform: Installation and Configuration Guide and any installation documentation for your particular solution.

If your system already has SAS software and you are adding a SAS solution, this topic steps you through using the SAS Deployment Wizard to create a SAS Application Server for a SAS solution. (SAS Demand-Driven Forecasting is the solution used to illustrate this procedure. However, the process is similar for most SAS solutions.)

To create a SAS Application Server for a SAS Solution, follow these steps:

Stop all running SAS servers and applications.

For more information, see Operating Your Servers in SAS Intelligence Platform: System Administration Guide.

Start the SAS Deployment Wizard from the highest-level directory in your SAS Software Depot using the command appropriate for your operating system:

Start-up Commands for the SAS Deployment Wizard

Operating System	Command
Windows	`setup.exe -record -deploy`
UNIX	`setup.sh -record -deploy`
z/OS	First invocation of the deployment wizard to install the metadata server as a 64-bit application: `setup.rexx -z64 -record -deploy` Second invocation of the deployment wizard to install the remainder of your SAS 9.3 order as a 31-bit application: `setup.rexx -record -deploy` To use a temporary directory other than `/tmp`, specify `-templocation` new-temp-dir.

Note: Using the

-record
                                 -deploy

options causes the wizard to create a response file that records the inputs that you supplied. This can be helpful when you want to repeat the deployment on other machines or when troubleshooting a deployment issue.

By default, the deployment wizard writes the response file in the following location:

Windows:

C:\Documents and Settings\current-user\sdwresponse.properties
UNIX and z/OS:

~/sdwresponse.properties

Choose Perform a Planned Deployment, and Install SAS Software only.
Next, select the deployment plan provided with the image.
Select the machine.

In this example, the server machine is used. Continue through the wizard. When you reach the Deployment Summary, click Start to begin the installation process. When the SAS Deployment Wizard finishes, your SAS solution and its dependent components are installed on the selected Server machine, and are ready to be configured. Restart all SAS servers and spawners, and then invoke the SAS Deployment Wizard a second time.
Select Perform a Planned Deployment, and specify Configure SAS Software only.
Next, specify the deployment plan provided with the image.

Select the machine. (In this example, the server machine is used.)
When prompted, specify the same configuration directory and level that is currently in use.

The deployment wizard recognizes that this configuration directory and level directory already contains files.
Click Yes to proceed.

A page listing the products to configure opens.

By default, the SAS Deployment Wizard presents only products that have not been configured. Any products required by your solution but not yet configured are automatically selected. However, since the deployment wizard is creating a new SAS Application Server, the SAS servers required and used by your solution remain to be configured.

In this example, the deployment wizard automatically specifies a SAS Pooled Workspace Server, a SAS Workspace Server, and a SAS Stored Process Server, to provide the required “exclusive” access to a SAS Application Server for your solution.
Select Next to proceed.

You must specify the same values that you used during the initial deployment here. List the host information for the SAS Metadata Server and valid credentials for the SAS Administrator, the SAS Trusted User, and SAS Spawned Servers account.

Continue through the wizard pages, specifying the same connection information until the Create SAS Application Server Context page appears.
Select Create a SAS Application Server Context.

A page prompts you for the name of the SAS Application Server. A unique name is required, and SAS suggests that a name easily associated with your solution be used. In this example, SASDDCF is specified.
Enter the name of your SAS Application Server context, and click Next.

For each server that you add to the SAS Application Server, you are prompted for the context to use. The same name that is used in the server context page should be used with each server.
Enter your SAS Application Server context and click Next.

The deployment wizard prompts you for a main port for each server that you add. The SAS Workspace Server and SAS Pooled Workspace Server being added to your server context can use whatever port you want, including the same port the previous configuration might have used for them (provided that no other conflicts exist). The SAS 9.3 object spawner manages these workspace servers to avoid port conflicts when it spawns them. However, the SAS Stored Process Server requires unique port values for its multi-bridge connections.

This example uses a variation on the default values of 8611, 8621, and 8631 combined with calculated defaults for ports not reserved. Any ports that are appropriate and available can be used. They merely must be unique and not in use.
Enter multi-bridge ports for the stored process server and click Next.

Continue through the wizard, providing input for the various configuration options as required by the solution being configured.
Specify the dependency for your solution server configuration, using the newly created SAS Application Server.

(In this example, the SAS Application Server is named SASDDCF.)

Continue through the wizard, providing configuration input.
When the Deployment Summary appears, click Start to begin the configuration process.

When the SAS Deployment Wizard finishes, your SAS solutions are installed and configured on the selected server machine and are ready to use. The wizard displays Configuration Guidelines and Details at the end of your deployment process.
Restart the SAS Object Spawner to detect the new servers that you have just deployed.

For more information, see Using Scripts to Operate SAS Servers Individually in SAS Intelligence Platform: System Administration Guide.