Job Scheduling

The Axelor Open Platform integrates a feature rich quartz scheduler library for job scheduling.

Jobs are executed by a separate process running under an admin session. The services executed can be transactional. All database operations will be executed using the admin access.

Configuration

The scheduler is disabled by default. It can be enabled with the following configuration:

# Quartz Scheduler
# ~~~~~
# quartz job scheduler

# Specify whether to enable quartz scheduler
quartz.enable = false

# total number of threads in quartz thread pool
# the number of jobs that can run simultaneously
quartz.thread-count = 3

JobStore

When the scheduler is enabled, RAMJobStore is used by default, which stores job and trigger information within memory. For persistence, you can configure Quartz to use a JDBC JobStore, which stores that information in the database. This is required for multi-instance deployments, but can also be useful for single-instance deployments to ensure missed jobs during downtime are executed at the next application startup.

The Axelor Open Platform automatically configures the JDBC JobStore to use the application’s main database and creates the necessary Quartz tables (qrtz_*) if they don’t exist on application startup.

To enable persistence using the JobStoreTX, set the following property:

# Use JobStoreTX for persistence instead of default RAMJobStore
quartz.job-store.class = org.quartz.impl.jdbcjobstore.JobStoreTX
Any extra quartz.job-store.* properties are passed and used to configure the job store via reflection in case you need advanced configuration. If you do that, make sure the appropriate setters exist on the job store.

JDBC JobStore Clustering

Quartz JobStoreTX job store with clustering features are automatically enabled when setting application.cache.provider to a distributed cache provider such as redisson or redisson-native. This is required for proper scheduling in multi-instance deployments.

Quartz clustering is further tunable with the following optional property:

# Frequency (in milliseconds) at which an instance "checks-in" with the other instances of the cluster
quartz.job-store.cluster-checkin-interval = 7500

JDBC JobStore Data Source

When using a JDBC JobStore (JobStoreTX whether clustered or not), an underlying HikariDataSource connection pool is configured to use the application’s main database settings and can be tuned with these optional properties:

# Maximum number of connections that the DataSource can create in its pool of connections (defaults to `quartz.thread-count`)
quartz.data-source.maximum-pool-size = 3

# Optional SQL query string that the DataSource can use to detect and replace failed/corrupt connections
quartz.data-source.connection-test-query =

# Discards connections after they have been idle this many milliseconds (disabled by default)
quartz.data-source.idle-timeout = 0
All quartz.data-source.* properties are passed and used to configure the data source via reflection in case you need advanced configuration. If you do that, make sure the appropriate setters exist on the data source.

Jobs

Scheduled jobs can be configured from Administration → Jobs → Schedules menu. The schedule configuration data requires:

  • name - name of the job

  • job - the job class implementing org.quartz.Job interface

  • cron - the cron string to schedule the job

  • active - whether the job is enabled

Additionally, job configuration can have parameter values (list of key → value pairs).

Here is an example job implementation:

package com.axelor.contact.jobs;

import org.quartz.Job;
import org.quartz.JobDataMap;
import org.quartz.JobDetail;
import org.quartz.JobExecutionContext;
import org.quartz.JobExecutionException;

public class HelloJob implements Job {

  @Override
  public void execute(JobExecutionContext context) throws JobExecutionException {
    JobDetail detail = context.getJobDetail();
    JobDataMap data = context.getJobDetail().getJobDataMap();

    String name = detail.getKey().getName();
    String desc = detail.getDescription();

    System.err.println("Job fired: " + name + " (" + desc + ")");
    if (data != null && data.size() > 0) {
      for (String key : data.keySet()) {
        System.err.println("    " + key + " = " + data.getString(key));
      }
    }
  }
}