Enterprise PeopleTools 8.51 PeopleBook: PeopleSoft Process Scheduler

Setting Server Definitions

This chapter provides an overview of server definitions and discusses how to:

Create server definitions.
Define report nodes.
Define daemon groups.
Define batch timings.

Understanding Server Definitions

Server definitions are created to help balance your workload on the system by identifying certain servers to perform processes. You determine which processes you want to schedule through PeopleSoft Process Scheduler and identify servers that you want slated to run specific types of processes. For example, you might want to have one server called RPTSRV to manage all reports and another to manage all journal posting.

To ensure that jobs continue processing without interruption and that they run on time after a hardware or software failure hosting a PeopleSoft Process Scheduler Server Agent, you can automatically reassign requests to an available agent based on the value that is selected for the Redistribute Workload option.

Creating Server Definitions

This section discusses how to:

Define servers.
Set distribution options.
Set operation times.
Set notification options.
Set daemon process options.
Monitor a daemon process.

Pages Used to Create Server Definitions

Page Name	Definition Name	Navigation	Usage
Server Definition	SERVERDEFN	PeopleTools, Process Scheduler, Servers, Server Definition	Define server agent options.
Distribution	SERVERCONTENT	PeopleTools, Process Scheduler, Servers, Distribution	Set distribution options.
Operation	SERVEROPRTN	PeopleTools, Process Scheduler, Servers, Operation	Set operation times.
Notification	SERVERNOTIFY	PeopleTools, Process Scheduler, Servers, Notification	Set notification options.
Daemon	SERVERDAEMON	PeopleTools, Process Scheduler, Servers, Daemon	Set daemon process options.

Defining Servers

To access the Server Definition page, select PeopleTools, Process Scheduler, Servers, Server Definition.

Sleep Time

Enter a sleep time, in seconds, for this server.

The PeopleSoft Process Scheduler Server Agent should not run continuously. To control its activity, you can schedule a sleep time, which refers to the number of seconds that you want the agent to wait, or sleep, before it checks for queued process requests in the Process Request table. When it wakes, it checks to see whether any processes have been queued in the Process Request table and need to be run on this process server.

For example, if you set the sleep time to 15 seconds and no process is queued, it wakes every 15 seconds and checks for queued processes. If it finds some, it processes as much as possible in 15 seconds and then goes back to sleep. If the work is not completed, it continues from the point at which it stopped for the next 15 seconds, and then goes back to sleep. The sleeping and polling process continues until a database or server administrator manually shuts down the agent.

Note. Depending on the server platform, you typically do not set the sleep time at fewer than 10 seconds. Between 15 and 30 seconds is generally recommended for most PeopleSoft applications. The maximum sleep time is 9,999 seconds (about 2 hours and 26 minutes).

Heartbeat

Enter a time interval, in seconds, for issuing a heartbeat message.

The PeopleSoft Process Scheduler Server Agent uses this value to track server status—running, down, or suspended. Each time the server issues a heartbeat message, it updates the last date and time stamp in the Server Status table with the current date and time. This prevents the database from accepting more than one PeopleSoft Process Scheduler Server Agent with the same name.

Max API Aware (maximum application programming interface aware tasks)

Enter the maximum number of API-aware tasks that can run concurrently on this server. An API-aware task is a process that properly updates its process status through the type-specific API that is provided, such as SQR, COBOL, and Crystal. It is the responsibility of the application process to update the Process Request table with status information.

Note. PSJob is not a valid process type to be considered in Max Concurrency and Process Type priority. PSJob is a container of processes and other PSJobs.

Operating System

An error message appears if you attempt to start the server agent on an operating system that is different from the operating system specified here.

CPU Utilization Threshold (%)

Enter a percentage threshold. If the amount of CPU utilization exceeds this threshold, the scheduler's status will change to Overload and it will wait to launch any new processes.

The Process Monitor - Server List page displays the current CPU utilization and the threshold value that is entered here.

Memory Utilization Threshold (%)

Enter a percentage threshold. If the amount of memory utilization exceeds this threshold, the scheduler's status will change to Overload and it will wait to launch any new processes.

The Process Monitor - Server List page displays the current memory utilization and the threshold value that is entered here.

Server Load Balancing Option

Select Use Option or Do Not Use Option. When you select Do Not Use Option, the server takes a request only when the name of the server has been specified on the Process Request page or in any of the definition tables.

The distributor does not assign a request to this server when the server name in the request is blank.

Redistribute Workload Option

Select Do Not Redistribute, Redistribute to any O/S, or Redistribute to Same O/S. When you select the option to redistribute, another active agent can take a request that is originally assigned to this server.

When this agent detects that one of the other active agents is no longer active and the server is set up to allow work to be distributed, it takes any queued request that was assigned to the inactive agent and redistributes it, based on the available active agents.

Note. After creating the server definition, you must configure the PeopleSoft Process Scheduler Server using PSADMIN.

Process Categories run on this Server

Priority	Select High, Medium or Low to prioritize all processes belonging to the corresponding process category that are queued to run on a server.
Max Concurrent	Enter the maximum number of the processes belonging to the corresponding process category that can run concurrently on this server. Max Concurrent is similar to Max API Aware, except that it controls how many processes of a process class can run concurrently on the server.

Warning! The processes that are contained in the categories must be of a type that are listed to run on the server. If the process type of a process is not listed, the process will not run.

Note. To disable a process category on this server, set the Max Concurrent value to 0.

Note. Parallel LoadCache has been assigned to the LOADCACHE process category. That means it will run only on servers that have a LOADCACHE process category whose MaxConcurrent value is greater than 0.

See Defining Process Categories.

Process Types Run on this Server

Process Types

Select the process types that the server should process. This selection enables server load balancing because you can direct particular processes to a specific server.

Priority

Select High, Medium, or Low to prioritize all processes that are queued to run on a server.

Max Concurrent

Enter the maximum number of the corresponding process class that can run concurrently on this server. Max Concurrent is similar to Max API Aware, except that it controls how many processes of a process class can run concurrently on the server.

Note. Each Winword instance that is initiated shares the common Winword template NORMAL.DOT. Therefore, some instances may encounter an error message when multiple Winword instances are initiated. The Max Concurrent value for the process type Winword should be set to 1 as a precautionary step to prevent this problem.

Important! Do not use the Max Concurrent field on the Server Definition page to specify the maximum number of concurrent processes for a PSJOB. To configure the maximum concurrent parameter for any given job, use the Max Concurrent field on the Job Definition page.

See Also

Using Process Request APIs

PeopleTools 8.51 Installation Guide for your database platform.

Setting Distribution Options

To access the Distribution page, select PeopleTools, Process Scheduler, Servers, Distribution.

Use the Distribution page to change the settings that the server uses to transfer output to Report Manager.

Distribution Node Name	Select the name of the report node.
Maximum Transfer Retries	Enter the number of times that the server can try to send a report to Report Manager before it quits.
Interval for Transfer Attempt	Enter the number of seconds that must pass before the server tries to transfer the report again.
Transfer System Files to Report Repository	Select to transfer system files to the Report Repository. For a non-Web output type, if this check box is cleared the link to system files on the report list page will be disabled.

Setting Operation Times

To access the Operation page, select PeopleTools, Process Scheduler, Servers, Operation.

Specify the days and times during which the server is operational. The preceding sample page shows a server that is operational 24 hours a day and seven days a week.

Setting Notification Options

To access the Notification page, select PeopleTools, Process Scheduler, Servers, Notification.

Use the Notification page to send messages to a group (role ID) or individuals (user ID) when an activity occurs with the server, such as an error or shutdown.

Limit overload notification to every n minutes	Enter the schedule for sending notification email when the server is overloaded.
ID Type	Select User or Role.
Distribution ID	Select the actual user ID or the group of users.
Server Errors	Select to notify the user or group when an error occurs. When this is selected, error notification for processes or jobs will be sent when a process or job goes to Error. This setting does not generate success or warning notifications. Success or warning notifications have to be individually configured in the process or job definition.
Down	Select to notify the user or group when the server shuts down.
Started	Select to notify the user or group when the server is started.
Suspended/Overloaded	Select to notify the user or group when activity on this server is suspended or overloaded.
Disabled	Select if you do not want to send notifications to the user or group.

Server Notification Examples

If you enter an ID Type and Distribution ID, then—depending on the check boxes selected—notifications will be sent to the appropriate roles or users. Notifications are sent based on the Limit overload notification to every n minutes value. If this field is left blank, the process scheduler server sends a notification every cycle. If a value is entered, the notification is sent every n minutes. This table provides examples of the notifications sent based on the check boxes selected:

Check Box Selected	Notification Sent When	Notification Message
Server Errors	A process by the server goes to error.	Message example: The Process Instance (252 - Process (XRFWIN) failed to complete successfully)
Down	Process Scheduler is shut down.	The Process Scheduler Server Agent <process scheduler server name> for database <database name> was shut down in server <machine name>.
Started	Process Scheduler is started.	The Process Scheduler Server Agent <process scheduler server name> for database <database name> was started in server <machine name>.
Suspended/Overloaded	Either CPU Utilization Threshold or Memory Utilization (as specified in Server Definition page) is exceeded.	Process Scheduler Server Agent <process scheduler server name> has exceeded the CPU Threshold.
Disabled	No notification is sent to this Distribution ID.	Not applicable.

Setting Daemon Process Options

To access the Daemon page, select PeopleTools, Process Scheduler, Servers, Daemon.

A daemon process is an Application Engine process that runs continuously when PeopleSoft Process Scheduler is operational. It triggers other application engine processes based on the daemon group that is entered.

Daemon Sleep Time	Schedule a sleep time to control the activity of the process. Because the daemon is a process that runs in the background, it should not run continuously. A sleep time is the number of minutes for which the daemon process sleeps, or waits, before it checks for work. When it wakes, it checks for processes that have been and need to be run on this process server.
Recycle Count	A cycle is the sequence of sleeping and working. The system automatically counts the number of times that it sleeps and works. When it reaches the recycle count value, the daemon process reboots itself.

Monitoring a Daemon Process

To access the Sever Details page, select PeopleTools, Process Scheduler, Process Monitor, Server List, Details.

Use the Process Monitor to monitor messages that are issued directly by the daemon and messages that are issued by the application engine programs that the daemon initiates.

To monitor a daemon process:

Select PeopleTools, Process Scheduler, Process Monitor.
Select the Server List page.
Click the Details link that is associated with the required process.

The Server Detail page appears, displaying information about the server and daemon group.
Click the Message Log link, located in the Daemon group box.

The Message Log page appears. A Delete button is located on this page, but it remains hidden when the daemon is running. When the Delete button appears, click it to delete all of the messages in the log.
Click the Return button to return to the Server Detail page.
Click the Cancel button to return to the Server List page.

Click the OK button if you have stopped, suspended, or restarted the server.

See Also

Viewing Server Details

Defining Report Nodes

This section provides an overview of report distribution nodes, lists common elements, and discusses how to:

Define HTTP distribution nodes.
Define File Transfer Protocol (FTP) distribution nodes.
Define XCopy distribution nodes.

Understanding Report Distribution Nodes

The report distribution node defines how your reports are moved to the Report Repository, where you can view them from Report Manager. Reports are moved using XCopy, FTP, or HTTP and HTTPS, depending on the type of server that you are using.

Before transferring the files to the Report Repository, determine which transfer protocol to use.

Setup	Transfer Protocol
Both PeopleSoft Process Scheduler and Web server on Microsoft Windows.	Use XCopy, FTP, or HTTP/HTTPS. If FTP information is not specified, PeopleSoft Process Scheduler performs an XCopy. Note. For XCopy, the Report Repository directory must be a shared drive on the network.
PeopleSoft Process Scheduler on Microsoft Windows and a UNIX Web server.	Use FTP or HTTP and HTTPS.
PeopleSoft Process Scheduler on OS390.	Use FTP or HTTP and HTTPS.

Note. If you are using FTP, the FTP daemon must be set up on the Web server.

Behavior of Shared Fields

The following fields are shared between the Http Distribution Node page and the FTP/XCopy Distribution Node page:

URL
Description
Operating System
Login ID
Password
Confirm Password

When you enter information on one page, the information also appears in the shared fields on the other page, but the fields are unavailable.

If you complete the information for one protocol and then change your selection to another protocol, the shared fields become active on the other page and unavailable on the original page. Upon saving, the system automatically clears fields that are not shared.

Common Element Used in This Section

URL

Enter the URL of the Web server with this format:

http://<machine name>:<port number>/psreports/<PeopleSoft site name>

Replace <machine name> with the name of the machine.

Pages Used to Define Report Nodes

Page Name	Definition Name	Navigation	Usage
Http Distribution Node	PRCSDISTNODE2	PeopleTools, Process Scheduler, Report Nodes, Http Distribution Node	Enter report node information to transfer reports to the Report Repository using http transfer protocol.
FTP/XCopy Distribution Node	PRCSDISTNODE	PeopleTools, Process Scheduler, Report Nodes, FTP/XCopy Distribution Node	Enter report node information to transfer reports to the Report Repository using FTP/XCopy transfer protocol.

Defining HTTP Distribution Nodes

To access the Http Distribution Node page, select PeopleTools, Process Scheduler, Report Nodes, Http Distribution Node.

Note. You should use web server basic authentication when you configure the SchedulerTransfer servlet that is used by the report distribution system. Please see the install guide for details.

http and https

Select the type of node connection. The default is http.

URI Host

Enter the machine name. If you are using PeopleSoft Process Scheduler for UNIX or OS390, you must enter either the fully qualified name (for example, ADHP04.peoplesoft.com or the IP address.

Warning! If you specify the Auth Token Domain name during the PeopleSoft Pure Internet Architecture installation, you must include a fully qualified domain name on the URL Host instead of the IP address. Otherwise, the Distribution Agent will not pass authentication.

URI Port

Enter the port number. This value must match the port number of the Web server. The defaults are 80 for HTTP and 443 for HTTPS.

Note. If you change a port number, you lose the default values for both protocols.

URI Resource

Enter SchedulerTransfer/<PeopleSoft site name>.

Login ID, Password, and Confirm Password

These are required only when the Web administrator has set up basic authentication on the Web server.

Save

Click to save your entries. To add additional distribution nodes, click the Add button to return to the search page.

File Chunking

The Distribution Agent automatically breaks up a large file and sends it in multiple HTTP posts. For example, a 150-MB file can be sent in 10 × 15 MB, 15 × 10MB, and so on.

To accommodate different hardware configurations (memory) and dissimilar Java Virtual Machine (JVM) tuning, two new parameters have been added to the PeopleSoft Process Scheduler section of the configuration file (prcs.cfg):

Chunking Threshold
File Chunk Size

These parameters enable clients to determine the most favorable configuration for their systems by trading off between the number of hits to the Web server when sending small chunks and memory usage due to sending large chunks.

See Also

Managing PeopleSoft Process Scheduler

Defining FTP Distribution Nodes

To access the FTP/XCopy Distribution Node page, select PeopleTools, Process Scheduler, Report Nodes, FTP/XCopy Distribution Node.

Home Directory

Enter the directory that is specified during the installation of PeopleSoft Pure Internet Architecture as the Report Repository.

Note. The FTP user ID must have write-access to this directory.

FTP Address

Enter the machine name or TCP/IP information for the Report Repository.

FTP ID, and Password

Enter the FTP user ID and password.

Defining XCopy Distribution Nodes

Access the FTP/XCopy Distribution Node page.

Home Directory, FTP Address, Password, FTP ID, and Confirm Password

Leave blank.

Network Path

Enter the universal naming convention (UNC) path that points to the Report Repository, for example:

\\<machine name>\psreports

Defining Daemon Groups

Use the Daemon Group page to enable a daemon process for the selected server.

To define a daemon group:

Select PeopleTools, Process Scheduler, Daemon Group.
Select the Add New Value page.
Enter a new daemon procedure group name.
Click the Add button.

The Daemon Group page appears.
Click the Load All Programs button to load all available application engine programs, or select a program to add.

Programs must be marked as daemon in their properties to be available.
Click the Save button.

Defining Batch Timings

This section provides an overview of batch timings and discusses how to define batch timings.

Understanding Batch Timings

Batch Timings reports are provided so that you can monitor the performance of your application engine programs. The Process Scheduler - Batch Timings page applies to the Statement Timings data that is stored in the (table) option.

For the Batch Timings feature to record data, you must enable it using the Configuration Manager Trace tab. In the Application Engine group, select the Statement Timings (table) option.

Note. Whenever you run an application engine program and the Statement Timings trace options are enabled, you can always view the batch timings results using the Process Monitor.

See Also

Using the PSADMIN Utility

Page Used to Define Batch Timings

Page Name	Definition Name	Navigation	Usage
Batch Timings	BANRUNCNTL	PeopleTools, Process Scheduler, Batch Timings	Monitor the performance of application engine programs.

Defining Batch Timings

To access the Batch Timings page, select PeopleTools, Process Scheduler, Batch Timings.

Run Control ID

Displays the run control ID that is used to run the Batch Timings report.

Report Type

Select the type of report that you want to generate, based on the data that is stored in the batch timings table.

Summary: Provides a report of all of the runs that are initiated by a run control ID. The Process Instance field is unavailable when this option is selected.

Detail: Provides a report of a specific run or process instance of an application engine program. The Run Control ID field is unavailable when this option is selected.

Batch Timings For

Based on the report type that you selected, enter the run control ID or process instance.

Sleep Time	Enter a sleep time, in seconds, for this server. The PeopleSoft Process Scheduler Server Agent should not run continuously. To control its activity, you can schedule a sleep time, which refers to the number of seconds that you want the agent to wait, or sleep, before it checks for queued process requests in the Process Request table. When it wakes, it checks to see whether any processes have been queued in the Process Request table and need to be run on this process server. For example, if you set the sleep time to 15 seconds and no process is queued, it wakes every 15 seconds and checks for queued processes. If it finds some, it processes as much as possible in 15 seconds and then goes back to sleep. If the work is not completed, it continues from the point at which it stopped for the next 15 seconds, and then goes back to sleep. The sleeping and polling process continues until a database or server administrator manually shuts down the agent. Note. Depending on the server platform, you typically do not set the sleep time at fewer than 10 seconds. Between 15 and 30 seconds is generally recommended for most PeopleSoft applications. The maximum sleep time is 9,999 seconds (about 2 hours and 26 minutes).
Heartbeat	Enter a time interval, in seconds, for issuing a heartbeat message. The PeopleSoft Process Scheduler Server Agent uses this value to track server status—running, down, or suspended. Each time the server issues a heartbeat message, it updates the last date and time stamp in the Server Status table with the current date and time. This prevents the database from accepting more than one PeopleSoft Process Scheduler Server Agent with the same name.
Max API Aware (maximum application programming interface aware tasks)	Enter the maximum number of API-aware tasks that can run concurrently on this server. An API-aware task is a process that properly updates its process status through the type-specific API that is provided, such as SQR, COBOL, and Crystal. It is the responsibility of the application process to update the Process Request table with status information. Note. PSJob is not a valid process type to be considered in Max Concurrency and Process Type priority. PSJob is a container of processes and other PSJobs.
Operating System	An error message appears if you attempt to start the server agent on an operating system that is different from the operating system specified here.
CPU Utilization Threshold (%)	Enter a percentage threshold. If the amount of CPU utilization exceeds this threshold, the scheduler's status will change to Overload and it will wait to launch any new processes. The Process Monitor - Server List page displays the current CPU utilization and the threshold value that is entered here.
Memory Utilization Threshold (%)	Enter a percentage threshold. If the amount of memory utilization exceeds this threshold, the scheduler's status will change to Overload and it will wait to launch any new processes. The Process Monitor - Server List page displays the current memory utilization and the threshold value that is entered here.
Server Load Balancing Option	Select Use Option or Do Not Use Option. When you select Do Not Use Option, the server takes a request only when the name of the server has been specified on the Process Request page or in any of the definition tables. The distributor does not assign a request to this server when the server name in the request is blank.
Redistribute Workload Option	Select Do Not Redistribute, Redistribute to any O/S, or Redistribute to Same O/S. When you select the option to redistribute, another active agent can take a request that is originally assigned to this server. When this agent detects that one of the other active agents is no longer active and the server is set up to allow work to be distributed, it takes any queued request that was assigned to the inactive agent and redistributes it, based on the available active agents.

http and https	Select the type of node connection. The default is http.
URI Host	Enter the machine name. If you are using PeopleSoft Process Scheduler for UNIX or OS390, you must enter either the fully qualified name (for example, ADHP04.peoplesoft.com or the IP address. Warning! If you specify the Auth Token Domain name during the PeopleSoft Pure Internet Architecture installation, you must include a fully qualified domain name on the URL Host instead of the IP address. Otherwise, the Distribution Agent will not pass authentication.
URI Port	Enter the port number. This value must match the port number of the Web server. The defaults are 80 for HTTP and 443 for HTTPS. Note. If you change a port number, you lose the default values for both protocols.
URI Resource	Enter SchedulerTransfer/<PeopleSoft site name>.
Login ID, Password, and Confirm Password	These are required only when the Web administrator has set up basic authentication on the Web server.
Save	Click to save your entries. To add additional distribution nodes, click the Add button to return to the search page.