Scheduling

In several web applications there is the need of running periodic tasks on the system. Bennu provides a simple way of doing so. To create tasks that can be scheduled the developer needs to implement a CronTask and annotate it with @Task, like so:

@Task(englishTitle = "Logger test")
public class LoggingTask extends CronTask {
    @Override
    public void runTask() {
        DateTime now = new DateTime();
        DateTime timestamp;
        do {
            timestamp = new DateTime();
            taskLog(timestamp.toString());
            try {
                Thread.sleep(1000);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        } while (timestamp.isBefore(now.plusMinutes(1)));
    }
}

By default the task runs in write mode, to make it read-only change the annotation to:

@Task(englishTitle = "Logger test", readOnly = true)

Inside the runTask() you can call methods annotated with @Atomic(mode=TxMode.WRITE) as in any other piece of code.

Configuring a task to run periodically is done in the application’s interface specifying a crontab like schedule periodicity (Cron)

Alternatively to developing a task in the codebase there is the possibility of submitting your code directly in the interface to run immediatly. To do so extendCustomTask instead, without annotation:

public class LoggingTask extends CustomTask {
    @Override
    public void runTask() {
        ...
    }
}

And use the Custom Task submission interface.

Logging

Every time a task is run, the Scheduler system keeps track of it, storing the log for the execution. In the log, the following are stored for future reference:

• Execution ID (random string assigned to each execution)
• Start Time
• End Time (only if the task has already finished)
• State (Either Running, Success or Failure)
• Task Name (fully qualified class name)
• Stack Trace (if an exception was thrown)
• File list (of files explicitly output by the task)
• Hostname for the application server that ran the task
• Code (only for custom tasks)
• Username of the user who ran the task (only for custom tasks)

Due to the potentially high volume of task execution, the number of log entries may increase really quickly.

Log Repository

The Scheduler provides an extension point to customize how logs are stored. The default implementation stores the logs in a file-system tree, keeping an index of the latest execution for each task in the tree’s root.