Distribution Server Features

Distribution Server Startup

In this section:

Configuration Considerations When the Distribution Server is Installed Separately From the TIBCO WebFOCUS Client

During startup, the Distribution Server calls the IBFS system to obtain information about how to communicate with the WebFOCUS Repository. During the WebFOCUS Client installation, you can specify an existing SQL data source or choose to have WebFOCUS install and configure the Apache Derby SQL data source. If the Distribution Server cannot communicate with the Repository, it will not start. If the Distribution Server does not start, check the Distribution Server startup trace and log files for error messages.

Upon successful initialization, the Distribution Server checks for jobs that need to be recovered, as specified by the Recovery and Scanback parameters and jobs that need to run as specified by the NEXTRUNTIME value of each schedule.

Configuration Considerations When the Distribution Server is Installed Separately From the TIBCO WebFOCUS Client

When the ReportCaster Distribution Server is installed as a stand-alone server, on a machine that is separate from the WebFOCUS Client, you must perform additional manual steps to ensure that configuration changes made through the WebFOCUS Administration Console are available to ReportCaster. This is because the stand-alone Distribution Server does not have access to the WebFOCUS configuration files being updated by the console. This is especially important when configuring external security for WebFOCUS, since your ReportCaster jobs may not run properly if the Distribution Server is not using the same security settings as the WebFOCUS Client.

The recommended procedure is to make changes to WebFOCUS, as per the documentation, and then test them with a web browser. When the settings are verified, perform the following steps to ensure that the configuration is available to ReportCaster:

Copy the webfocus.cfg file and the install.cfg file from the ..\ibi\WebFOCUS90\config directory to the ..\ibi\WebFOCUS90\config directory on the stand-alone Distribution Server machine.
Copy the odin.cfg file from the ..\ibi\WebFOCUS90\client\wfc\etc directory to the ..\ibi\WebFOCUS90\client\wfc\etc directory on the stand-alone Distribution Server machine.
Copy the cgivars1.wfs file from the ..\ibi\WebFOCUS90\client\wfc\etc directory to the ..\ibi\WebFOCUS90\client\wfc\etc directory on the stand-alone Distribution Server machine.
Restart the Distribution Server and test the scheduled job behavior.

Recovery

Reference:

How Jobs Are Recovered

The purpose of the Recovery parameter in the ReportCaster Server Configuration tool is to recover jobs that were placed in the Distribution Server queue, but did not complete schedule execution processing. This could be due to the Distribution Server or Reporting Server being stopped. You can activate the recovery feature by setting Recovery to ON in the ReportCaster Server Configuration tool.

When you create a schedule and the Recovery parameter is set to ON, the RECOVERY field for the schedule is set to N in the BOTSCHED repository table. When the schedule is placed in the run queue, the RECOVERY field is set to Y. This means that if the Distribution Server becomes unavailable while the job is still in the queue, ReportCaster will recover the job when the Distribution Server becomes available, even if the Recovery parameter is set to OFF in the ReportCaster Server Configuration tool. After the job is run and all log records have been written, the RECOVERY field is set back to N.

Any job that is running because the Recovery parameter is set to ON runs only once. After the job is placed in the run queue, its NEXTRUNTIME is updated to the next time it should run after the current time. For example, if a job is scheduled to run hourly and the Distribution Server is unavailable for four hours, when the Distribution Server becomes available, the job will run only once and will resume running on an hourly basis thereafter.

Note:

The Recovery option works independently of the Scanback option. For more information about the Scanback option, see Scanback.
Only scheduled jobs can be recovered. Jobs that were submitted on demand cannot be recovered.

Reference: How Jobs Are Recovered

Scheduled jobs are recovered as follows:

When you start the Distribution Server, it checks for Recovery ON in the ReportCaster configuration.
If Recovery is set to ON, the Distribution Server reads all records from the BOTSCHED table whose RECOVERY field is set to Y and places them in the run queue, regardless of schedule information and NEXTRUNTIME value.
The Distribution Server begins the process of polling and looking for jobs to run after having queued all the jobs whose RECOVERY field is set to Y.
After the scheduled job is placed on the Distribution Server queue, its NEXTRUNTIME is updated to the next time it should run after the current time.

Note:

If the Recovery parameter is set to OFF, the Distribution Server modifies all jobs in the BOTSCHED table with a RECOVERY field value of Y to N. This turns recovery off for all jobs.
If the Recovery parameter is set back to ON, it is only activated for jobs whose NEXTRUNTIME is later than the time at which the Recovery parameter was set back to ON.

Scanback

In certain instances, the Distribution Server may be unavailable for a period of time. During this time, no scheduled jobs can run. By default, when the Distribution Server becomes available again, it searches for and runs all jobs with a next run time less than the current time and resets the next run time according to the next scheduled run time of the job. If the Distribution Server is unavailable for a long time, you may want to adjust this behavior. Using the Scanback parameter, you can run only those jobs found in a specified time period of the outage, or just reset the next run time of all unexecuted jobs without running them.

There are two Scanback parameters:

Scanback Type (On, Off, or Reset Next Run Time)
Scanback Interval (number of days)

The Scanback Interval is active only when the Scanback Type is On. For more information, see Distribution Server Settings.

The Scanback Type parameter can have the following settings:

On, when Scanback interval is an integer value greater than zero. Represents the number of 24-hour periods (beginning with the Distribution Server restart time) that the Distribution Server will scan back to look for and run jobs with a next run time less than the current time.
For example, if the Distribution Server is unavailable for three days and Scanback is set to 2, then the Distribution Server will only run those jobs it finds with a next run time within the previous 48 hours of the server restart time.

A valid value is any integer from 1 to 365. The default value is 15 days.
Reset Next Run Time. Looks for all jobs with a next run time less than the current time and resets the next run time to the next scheduled run time of the job.
Off. Disables Scanback. This means the Distribution Server follows its default behavior to find, run, and reset all jobs with a next run time that is less than the current time.

Note:

The Scanback option works independently of the Recovery option. For more information about the Recovery option, see Recovery.
Consider Daylight Saving Time when setting the Scanback parameter. For more information, see http://webexhibits.org/daylightsaving/b.html.

ReportCaster Server Configuration Tool

The ReportCaster Server Configuration tool is an administration tool that allows authorized users the ability to navigate through and change the various settings that define the ReportCaster configuration. For example, the polling interval for the Distribution Server can be changed and the Maximum Threads to control the number of connections available for each Reporting Server can be defined. The configuration settings are accessed from the Configuration tab in the ReportCaster Console.

For more information about ReportCaster configuration settings, see Configuring ReportCaster.

Distribution Server Failover and Workload Distribution

ReportCaster provides the ability to incorporate a backup Distribution Server, and to configure two or more Distribution Servers to share the work of executing scheduled jobs.

The Distribution Server Failover feature enables you to configure a backup Distribution Server that can resume ReportCaster operations when there is an interruption (planned or unplanned) in the primary Distribution Server service. A synchronized service monitors the primary server to verify that it is operational. If there is an interruption in service, the monitor triggers the failover server to take over the role of the primary server. For details on setting up a failover Distribution Server, see the WebFOCUS ReportCaster Post-Installation Tasks chapter in the TIBCO WebFOCUS® Installation and Configuration for Windows manual for your platform.
The ReportCaster Console provides information, such as mode and status, on the primary and failover servers. From the ReportCaster Console, you can suspend, start, and stop the Distribution Servers.

Note: If you have the ReportCaster Server Configuration Recovery option set to On, the Distribution Server recovers scheduled jobs that were processed but not completed.
The Workload Distribution feature enables ReportCaster to distribute the scheduled jobs across multiple Distribution Servers, providing an efficient and fast way to process large numbers of ReportCaster schedules.
Multiple Distribution Server instances can be installed on one or more hosts. One instance is designated as the Workload Manager, while the others are designated as Workers. The WebFOCUS Repository will be shared by the Workload Manager and the Workers.

Note: To configure for Workload Distribution, use the ReportCaster Configuration tool settings under the Distribution Servers folder. All servers will share one set of configuration information. The Workload Manager will push any configuration changes to the Workers.

The Workload Manager will poll the WebFOCUS repository to retrieve jobs scheduled for execution and place the jobs in its queue. It will then dispatch individual jobs to the next available Worker.

The Workers will queue the jobs sent by the Workload Manager and dispatch the jobs for execution. The Workers will distribute output from the executed jobs and update the ReportCaster Log files.

The following diagram illustrates the processing and relationship between the Workload Manager and its Workers.

Distribution Server Failover can apply to the Workload Manager, which is the primary Distribution Server, but not to the Workers.

You can monitor and manage the Workload Manager and Workers through the ReportCaster Console, Server Status window.