 |
|
| Understanding Server Pooling |
How Server-side Pooling Works
The usage of the servers in a server-side pool is governed by the authorization rules that are set on the servers in the SAS metadata. Server-side pools use the ReadMetadata permission on the server component metadata objects in the pool to control access. Server-side pooling allows more flexibility than client-side pooling, which limits access to a single group.
Another benefit of server-side pooling is that workspace server load balancing functionality is automatically built in. When a user connects to a pool, the spawner determines the server to send the user to based on the user's access rights and on current server loads.
By using access controls on servers in metadata, the server that actually runs a user's job can be controlled by the user's identity in metadata. (A server defined in metadata, can be one or more processes that are running on one or more machines.) Because different servers can have different command lines, you can use server-side pooling to ensure that a user is assigned to a server process that was launched with a specific command line. One useful implementation of this feature is to launch server-side pooled workspace servers that use an operating system's run priority level.
For example, on UNIX, an administrator might manage server performance by creating the server with the UNIX nice command. Lower priority users might use a server that was launched with a higher system value such as 5. Higher priority users might use a server that was launched with a lower system value such as -2. On Windows, the administrator might manage server performance using the /<priority class> option on the start command. (A user with a lower priority might run on a pooled workspace server started with a start-up command that is similar to: start/low sas.exe. Higher priority users might run on a pooled workspace server with a start-up command that is similar to: start/high sas.exe.) For more information, refer to your documentation for the appropriate operating system.
|
|
Copyright © 2010 by SAS Institute Inc., Cary, NC, USA. All rights reserved.