Understanding the State of Your System

Overview of the SAS Metadata Server and Its Initial Configuration

The SAS Metadata Server is the most critical software component in the SAS Intelligence Platform. All of the client applications and the other SAS servers that are part of the platform depend on the SAS Metadata Server and cannot operate without it.

The SAS Metadata Server is a multi-user server that serves metadata from one or more SAS Metadata Repositories to all of the SAS Intelligence Platform client applications in your environment. The SAS Metadata Server enables centralized control so that all users access consistent and accurate data.

The functionality of the SAS Metadata Server is provided through the SAS Open Metadata Architecture, which is a metadata management facility that provides common metadata services to applications. One metadata server supports all of the applications in your environment and can support hundreds of concurrent users. This architecture enables the following:

the exchange of metadata between applications, so that applications can work together more easily.
centralized management of metadata resources. Because there is a common framework for creating, accessing, and updating metadata, it is easier to manage the applications that rely on this metadata.

The SAS Metadata Server stores information about the following:

the enterprise data sources and data structures that are accessed by SAS applications
resources that are created and used by SAS applications, including information maps, OLAP cubes, report definitions, stored process definitions, and scheduled jobs
the servers that run SAS processes
the users and groups of users that use the system, and the levels of access that users and groups have to resources

About the Metadata Manager Plug-in

In SAS Management Console, the SAS Metadata Server appears in the Metadata Manager plug-in, as shown here:

[Metadata Manager tree in SAS Management Console ]

You can use this plug-in to stop, pause, resume, or reset the metadata server; to check the status of the metadata server and the metadata repositories; and to create and manage metadata repositories. You can also access metadata utilities such as the Backup and Restore Wizard.

Note: In the Plug-ins view, the Properties action generally displays an object's metadata definition. However, when you select Properties on the Active Server node of the Metadata Manager plug-in, the information that is displayed is supplied directly by the metadata server. [cautionend]

About the SASMeta Server Context

So that you can make use of SAS Management Console's server monitoring features, the metadata server also appears in the SASMeta application server tree (also referred to as the SASMeta server context) in the Server Manager plug-in, as shown here:

[untitled graphic]

As shown in the preceding display, the SASMeta server context also includes a workspace server and a DATA step batch server. This workspace server is only visible to and accessible by users who are in the SAS Administrators group. This server is used to execute metadata utilities including the Backup and Restore Wizard and the Import SAS Package and Export SAS Package wizards.

CAUTION:: Use the SASMeta workspace server only for running the metadata utilities. Do not use it for other SAS processing.

The DATA step batch server in the SASMeta context is provided for scheduling backup jobs that are created by using the Backup Wizard.

About the SAS Metadata Repositories and the Repository Manager

The SAS Metadata Server is a SAS process that enables clients to read metadata from and write metadata to one or more SAS Metadata Repositories. A repository is a physical location in which a collection of related metadata objects is stored.

Metadata repositories are managed by a repository manager, which is a library of tables that define the set of repositories that can be accessed by a SAS Metadata Server. A metadata server cannot be started without a repository manager. Each metadata server can have only one repository manager.

The SAS Metadata Server supports three types of metadata repositories:

foundation repository

the required metadata store for a SAS Metadata Server. Each metadata server has one foundation repository. As a default, the name of the foundation repository is Foundation. (If you performed a custom installation, then you might have assigned a different name.)

project repository

a metadata repository that serves as an isolated work area or playpen. Project repositories are available for SAS Data Integration Studio only. In general, each user who participates in change management has his or her own project repository.

A project repository enables a user to check out and lock metadata from a foundation repository or a custom repository so that the metadata can be modified and tested in a separate area. When the metadata is ready, it is checked in and unlocked so that it is available to other users.

custom repository

an optional metadata store for a SAS Metadata Server, in addition to the foundation repository. Custom repositories are useful for physically segregating metadata for storage or security purposes.

Custom repositories appear as separate root folders in the Folders tab of SAS Management Console.

In SAS Management Console, the repositories are displayed in the Metadata Manager plug-in, as shown in this example:

[Metadata Manager tree in SAS Management Console, with repositories displayed]

In SAS 9.2, it is not necessary to define dependencies between repositories. If a repository dependency was defined in SAS 9.1.3, then the dependency relationship is eliminated during the SAS 9.1.3 to SAS 9.2 migration process. However, the existing associations between metadata objects still remain in effect.

How the SAS Metadata Server Works

SAS Intelligence Platform applications connect to the SAS Metadata Server and issue SAS Open Metadata Interface method calls that access SAS Metadata Repositories.

To enable high-speed access by clients, the SAS Metadata Server is an "in-memory" server. As clients submit queries and updates, the data sets that contain the requested records are read from repository data sets on disk into the server's memory, where they remain until the server is paused and resumed or until the server is stopped.

In the default configuration, metadata updates that are requested by client applications are recorded immediately in memory and in a journal file so that client access can be resumed. The more time-consuming updates to the repository data sets on disk are performed later in the background.

The Initial Configuration of the SAS Metadata Server

The SAS Deployment Wizard provides the following initial configuration for your metadata server:

Your metadata server contains a foundation metadata repository called Foundation. (If you performed a custom installation, then you might have assigned a different name.) This repository contains the following items:
- metadata definitions for initial users, groups, and roles
- default authorization settings that give the initial users, groups, and roles access to different levels of functionality
- metadata definitions for your servers
- a full set of resource templates that can be used to create additional metadata objects
The data sets that make up the foundation repository are typically located in the following path:

SAS-configuration-directory\Lev1\SASMeta\
MetadataServer\MetadataRepositories\Foundation

In this path, SAS-configuration-directory refers to the SAS configuration directory (see About the Configuration Directory).
Journaling has been enabled.

When journaling is not enabled, clients that request metadata updates must wait for updates to be written to the repository data sets on the file system before client access is resumed. When journaling is enabled, access is returned to clients as soon as the metadata updates are written to the in-memory database and the journal file. The more time-consuming updates to the repository data sets are performed later in the background.

The default settings for journaling are as follows:
- The name of the journal file is MetadataJournal.dat.
- The journal file is written to the following path:
  
  SAS-configuration-directory\Lev1\SASMeta\MetadataServer\Journal
- On 32-bit machines and z/OS systems, 200 megabytes of storage on the metadata server are reserved for the journal file and cannot be used for other processes. On 64-bit machines, the default setting is 500 megabytes. To modify these settings, see Configuring Metadata Server Journaling.
Two special users of the metadata server have been set up: the SAS Administrator and the SAS Trusted User.
The metadata server has been configured to write daily logs. The default location for logs is as follows:

SAS-configuration-directory\Lev1\SASMeta\MetadataServer\Logs

For details about the default logging configuration, see Initial Logging Configuration for SAS Servers.

Top of Page