The Scheduler is modeled after Unix Cron. The Scheduler runs as a background process that executes timed scheduled tasks independently of HTTP requests. Tasks are either stored in the database in the TURBINE_SCHEDULED_JOB table and once entered in the database are loaded automatically when Turbine initializes or as a non persistent service using TurbineNonPersistentSchedulerService.
For the Scheduler to load classes that extend the ScheduledJob Class, the scheduler needs to be enabled via the TurbineResources.properties file, where the directive services.SchedulerService.enabled needs to be set to true.
The Scheduler Service should be accessed in one of two ways.
# ------------------------------------------------------------------- # # S E R V I C E S # # ------------------------------------------------------------------- # Classes for Turbine Services should be defined here. # Format: services.[name].classname=[implementing class] # # To specify properties of a service use the following syntax: # service.[name].[property]=[value] services.SchedulerService.classname=org.apache.turbine.services.schedule.QuartzSchedulerService . . . # ------------------------------------------------------------------- # # S C H E D U L E R S E R V I C E # # ------------------------------------------------------------------- # # Set enabled to true to start the scheduler. The scheduler can be # stopped and started after Turbine has been intialized. See the # javadocs for org.apache.turbine.services.schedule.TurbineScheduler # for the methods calls. # # Default = false # services.SchedulerService.enabled=false # Determines if the scheduler service should be initialized early. This # Should always be set to true!!!! services.SchedulerService.earlyInit=true # ------------------------------------------------------------------- # # P U L L S E R V I C E # # ------------------------------------------------------------------- # These are the properties for the Pull Service, the service # that works in conjuction with the Turbine Pull Model API. # ------------------------------------------------------------------- # This is a tool that allows access to the scheduler service. tool.request.scheduler=org.apache.turbine.services.schedule.SchedulerTool
Jobs are stored and retrieved from the database using Torque generated objects. The objects are based on the definition stored in scheduler-schema.xml. If you want to add additional fields to track more information about your scheduled tasks, you can modify this file to do so. You will need to rebuild Turbine using MAven after making your modifications.
To create an object that takes advantage of the Schedule Service, the task to be executed will need to be embedded in the run() method of an object which extends the ScheduledJob Class. For instance:
package com.mycompany.modules.scheduledjobs;
//JDK
import java.util.Date;
//Turbine
import org.apache.turbine.modules.ScheduledJob;
import org.apache.turbine.services.schedule.JobEntry;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
public class SimpleScheduledTask extends ScheduledJob
{
/** Logging */
private static Log log = LogFactory.getLog(SimpleScheduledTask.class);
private int taskcount = 0;
/**
* Constructor
*/
public SimpleScheduledTask()
{
//do Task initialization here
}
/**
* Run the Jobentry from the scheduler queue.
* From ScheduledJob.
*
* @param job The job to run.
*/
public void run( JobEntry job ) throws Exception
{
log.note("Scheduled job " + job.getJobId() + " : " +
"task: " + job.getTask() +
" ran @: " +
new Date(System.currentTimeMillis()).toString() +
" taskcount " + taskcount
);
//iterate the task counter
taskcount++;
}
}
The SimpleScheduledTask object makes an entry into the turbine.log each time the Task is run by the Schedule Service. The JobEntry object carries the information the Service uses to determine the frequency of the Task. Note that the object is in the com.mycompany.modules. This assumes that the package com.mycompany.modules is in the Turbine module path, which is in the module.packages directive in the TurbineResources.properties. The ScheduledJobLoader object loads the Task upon Turbine initilization also searches for a scheduledjobs package which contains the Task object.
Control of the time between, or to, execution of the Task, is controlled by the JobEntry object. The JobEntry serves as a wrapper for the scheduled task. The constructor is as follows:
public JobEntry(int sec, int min, int hour,
int wd, int day_mo, String task)
throws Exception
The granularity of the Task's next execution can be controlled to the level of seconds. In the above constructor, sec represents the seconds with a valid range of 0-59, min represents the minutes with a valid range of 0-59, hour represents the hours with a valid range of 0-23, wd is the day of the week with a valid range of 1-7, day_mo is the day of the month with a valid range of 1-31 and task is the name of the object. The JobEntry constructor allows for the Task to be run at a certain point in time. For example:
JobEntry je = new JobEntry(0,25,-1,-1,-1,"SimpleScheduledTask"); o run every 25 minutes. JobEntry je = new JobEntry(0,0,8,-1,15,"SimpleScheduledTask"); o run at 8:00am on the 15th of the month every month.
The Schedule Service will only execute Tasks at Turbine Initialization that are present in the Database. To add a Task to the Database we need to set up an Action class that can be accessed from a template with:
$page.SetTitle("SimpleScheduleTask Starter Page")
Set Values for SimpleScheduleTask and then start it for the first time.
<br />
<form>
<input type="hidden" name="action" value="SchedulerStart" />
<input type="text" name="second" size="20" /> : seconds (0-59)<br />
<input type="text" name="minute" size="20" /> : minutes (0-59)<br />
<input type="text" name="hour" size="20" /> : hours (0-23)<br />
<input type="text" name="weekday" size="20" /> : Day of the Week (1-7)<br />
<input type="text" name="day_of_month" size="20" /> Day of the Month (1-31)<br />
<input type="text" name="task" value="SimpleSchedulerTask" /> : The Task being scheduled. <br />
<input type="text" name="email" /> : Email<br />
</form>
and the appropriate Action class:
package com.mycompany.modules.actions;
//Turbine
import org.apache.turbine.modules.actions.VelocityAction;
import org.apache.turbine.services.schedule.JobEntry;
import org.apache.turbine.services.schedule.TurbineScheduler;
import org.apache.turbine.services.TurbineServices;
import org.apache.turbine.util.ParameterParser;
import org.apache.turbine.util.RunData;
public class AddScheduledTask extends VelocityAction
{
public void doPerform(RunData data) throws Exception
{
ParameterParser params = data.getParameters();
int second = params.getInt("second",-1);
int minute = params.getInt("minute",-1);
int hour = params.getInt("hour",-1);
int weekday = params.getInt("weekday",-1);
int dom = params.getInt("day_of_month",-1);
String task = params.getString("task","");
String email = params.getString("email","");
try
{
//create a new JobEntry with the time constraints from
//the template as the arguments
JobEntry je = new JobEntry(second,minute,hour,weekday,dom,task);
//set the email for notification
je.setEmail(email);
//add the job to the queue
TurbineScheduler.addJob(je);
//set the Message
data.setMessage("Task " + task + " added successfully");
}
catch (Exception e)
{
//set the Message
data.setMessage("Task " + task + " could not be added!");
}
// Note: The template "SchedulerStatus.vm" is not part of Turbine.
// It is only here as an example how you might implement this action.
setTemplate(data, "SchedulerStatus.vm");
}
}
The AddScheduledTask action class adds the task to the job queue. The job will also be written to database so that it will be automatically loaded the next time that Turbine starts.
The AddScheduledTask action class is really only part of what you would need to implement in order to control the job scheduler from a web interface. It is given as a VERY simple example of how you might implement such functionality. In a more complete implementation, you would have templates to create/edit a scheduled task, display all tasks, and control the scheduler. You would also have action(s) to enable/disable the scheduler and add/update/remove scheduled tasks.
The TurbineScheduler class exposes methods that allow the Scheduler process to be monitored and controlled, such as getJob(int oid), removeJob(JobEntry je), updateJob(JobEntry je), listJobs(), startScheduler(), stopScheduler(), and others. See the javadocs on this class for more details.
There is also a pull tool for use with the scheduler service (assuming the pull service is enabled). This tool allows you to get basic information from the scheduler service for use in a Velocity template. It provides methods such as getJob(jobId), listJobs(), isEnabled(), and others. See the javadocs for more details.
The Scheduler Service uses a seperate thread for each Task it runs to ensure that every job runs on time. It's the programmer's responsibility to ensure that proper precautions to handle issues such as synchronization and long running jobs. As always, check through the relevant Javadocs and source code for more details on the Scheduler Service.