Setting the Status of a Project

Each project in Intelligence Server can operate in one of several modes. Project modes allow for various system administration tasks to occur without interrupting Intelligence Server operation for other projects. The tasks that are allowed to occur depend on the job or jobs that are required for that task.

A project's status can be one of the following:

• Loaded
• Unloaded
• Request Idle
• Execution Idle
• Warehouse Execution Idle
• Full Idle
• Partial Idle

For instructions on changing a project's status, see Changing the Status of a Project.

For example scenarios where the different project idle modes can help to support project and data warehouse maintenance tasks, see Project and Data Warehouse Maintenance Example Scenarios.

Loaded

A project in Loaded mode appears as an available project in Developer and MicroStrategy Web products. In this mode, user requests are accepted and processed as normal.

Unloaded

Unloaded projects are still registered on Intelligence Server, but they do not appear as available projects in Developer or MicroStrategy Web products, even for administrators. Nothing can be done in the project until it is loaded again.

Unloading a project can be helpful when an administrator has changed some project configuration settings that do not affect run-time execution and are to be applied to the project at a later time. The administrator can unload the project, and then reload the project when it is time to apply the project configuration settings.

A project unload request is fully processed only when all executing jobs for the project are complete.

Request Idle

Request Idle mode helps to achieve a graceful shutdown of the project rather than modifying a project from Loaded mode directly to Full Idle mode. In this mode, Intelligence Server:

• Stops accepting new user requests from the clients for the project.
• Completes jobs that are already being processed. If a user requested that results be sent to their History List, the results are available in their History List after the project is resumed.

Setting a project to Request Idle can be helpful to manage server load for projects on different clusters. For example, in a cluster with two nodes named Node1 and Node2, the administrator wants to redirect load temporarily to the project on Node2. The administrator must first set the project on Node1 to Request Idle. This allows existing requests to finish execution for the project on Node1, and then all new load is handled by the project on Node2.

Execution Idle

A project in Execution Idle mode is ideal for Intelligence Server maintenance because this mode restricts users in the project from running any job in Intelligence Server. In this mode, Intelligence Server:

• Stops executing all new and currently executing jobs and, in most cases, places them in the job queue. This includes jobs that require SQL to be submitted to the data warehouse and jobs that are executed in Intelligence Server, such as answering prompts.

  If a project is idled while Intelligence Server is in the process of fetching query results from the data warehouse for a job, that job is canceled instead of being placed in the job queue. When the project is resumed, if the job was sent to the user's History List, an error message is placed in the History List. The user can click the message to resubmit the job request.

• Allows users to continue to request jobs, but execution is not allowed and the jobs are placed in the job queue. Jobs in the job queue are displayed as "Waiting for project" in the Job Monitor. When the project is resumed, Intelligence Server resumes executing the jobs in the job queue.

  This mode allows you to perform maintenance tasks for the project. For example, you can still view the different project administration monitors, create reports, create attributes, and so on. However, tasks such as element browsing, exporting, and running reports that are not cached are not allowed.

Warehouse Execution Idle

A project in Warehouse Execution Idle mode is ideal for data warehouse maintenance because this mode restricts users in the project from running any SQL against the data warehouse. In this mode, Intelligence Server:

• Accepts new user requests from clients for the project, but it does not submit any SQL to the data warehouse.
• Stops any new or currently executing jobs that require SQL to be executed against the data warehouse and, in most cases, places them in the job queue. These jobs display as "Waiting for project" in the Job Monitor. When the project is resumed, Intelligence Server resumes executing the jobs in the job queue.

  If a project is idled while Intelligence Server is in the process of fetching query results from the data warehouse for a job, that job is canceled instead of being placed in the job queue. When the project is resumed, if the job was sent to the user's History List, an error message is placed in the History List. The user can click the message to resubmit the job request.

• Completes any jobs that do not require SQL to be executed against the data warehouse.

  This mode allows you to perform maintenance tasks on the data warehouse while users continue to access non-database-dependent functionality. For example, users can run cached reports, but they cannot drill if that drilling requires additional SQL to be submitted to the data warehouse. Users can also export reports and documents in the project.

Full Idle

Full Idle is a combination of Request Idle and Execution Idle. In this mode, Intelligence Server does not accept any new user requests and active requests are canceled. When the project is resumed, Intelligence Server does not resubmit the canceled jobs and it places an error message in the user's History List. The user can click the message to resubmit the request.

This mode allows you to stop all Intelligence Server and data warehouse processing for a project. However, the project still remains in Intelligence Server memory.

Partial Idle

Partial Idle is a combination of Request Idle and Warehouse Execution Idle. In this mode, Intelligence Server does not accept any new user requests. Any active requests that require SQL to be submitted to the data warehouse are queued until the project is resumed. All other active requests are completed.

This mode allows you to stop all Intelligence Server and data warehouse processing for a project, while not canceling jobs that do not require any warehouse processing. The project still remains in Intelligence Server memory.

Changing the Status of a Project

To Load or Unload a Project

If the project is running on multiple clustered Intelligence Servers, the project is loaded or unloaded from all nodes. To load or unload the project from specific nodes, use the Cluster view instead of the Project view. For detailed instructions, see Using the Cluster View.

1. In Developer, log in to the project source containing the project.
2. Under that project source, expand Administration, then expand System Administration, and select Project.
3. Right-click the project, point to Administer Project, and select Load or Unload. The project is loaded or unloaded. If you are using clustered Intelligence Servers, the project is loaded or unloaded for all nodes in the cluster.

To Idle or Resume a Project

If the project is running on multiple clustered Intelligence Servers, the project status changes for all nodes. To idle or resume the project on specific nodes, use the Cluster view instead of the Project view. For detailed instructions, see Using the Cluster View.

1. In Developer, log in to the project source containing the project.
2. Under that project source, expand Administration, then expand System Administration, and then select Project.
3. Right-click the project, point to Administer Project, and select Idle/Resume.

   

4. Select the options for the idle mode that you want to set the project to:
   • Request Idle (Request Idle): all executing and queued jobs finish executing, and any newly submitted jobs are rejected.
   • Execution Idle (Execution Idle for All Jobs): all executing, queued, and newly submitted jobs are placed in the queue, to be executed when the project resumes.
   • Warehouse Execution Idle (Execution Idle for Warehouse jobs): all executing, queued, and newly submitted jobs that require SQL to be submitted to the data warehouse are placed in the queue, to be executed when the project resumes. Any jobs that do not require SQL to be executed against the data warehouse are executed.
   • Full Idle (Request Idle and Execution Idle for All jobs): all executing and queued jobs are canceled, and any newly submitted jobs are rejected.
   • Partial Idle (Request Idle and Execution Idle for Warehouse jobs): all executing and queued jobs that do not submit SQL against the data warehouse are canceled, and any newly submitted jobs are rejected. Any currently executing and queued jobs that do not require SQL to be executed against the data warehouse are executed.

     To resume the project from a previously idled state, clear the Request Idle and Execution Idle check boxes.

5. Click OK. The Idle/Resume dialog box closes and the project goes into the selected mode. If you are using clustered Intelligence Servers, the project mode is changed for all nodes in the cluster.