原文地址: https://devcenter.heroku.com/articles/scheduled-jobs-custom-clock-processes-java-quartz-rabbitmq

Table of Contents

There are numerous ways to schedule background jobs in Java applications. This article will teach you how to setup a Java application that uses the Quartz library along with RabbitMQ to create a scalable and reliable method of scheduling background jobs on Heroku.

Many of the common methods for background processing in Java advocate running background jobs within the same application as the web tier. This approach has scalability and reliability constraints.

A better approach is to move background jobs into separate processes so that the web tier is distinctly separate from the background processing tier. This allows the web tier to be exclusively for handling web requests. The scheduling of jobs should also be it’s own tier that puts jobs onto a queue. The worker processing tier can then be scaled independently from the rest of the application.

If you have questions about Java on Heroku, consider discussing them in the Java on Heroku forums.

The source for this article's reference application is available on GitHub.

Prerequisites

To clone the sample project to your local computer run:

$ git clone https://github.com/heroku/devcenter-java-quartz-rabbitmq.git
Cloning into devcenter-java-quartz-rabbitmq... $ cd devcenter-java-quartz-rabbitmq/

Scheduling jobs with Quartz

custom clock process will be used to create jobs and add them to a queue. To setup a custom clock process use the Quartz library. In Maven the dependency is declared with:

See the reference app's pom.xml for the full Maven build definition.

<dependency>
<groupId>org.quartz-scheduler</groupId>
<artifactId>quartz</artifactId>
<version>2.1.5</version>
</dependency>

Now a Java application can be used to schedule jobs. Here is an example:

package com.heroku.devcenter;

import org.quartz.*;
import org.quartz.impl.StdSchedulerFactory;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory; import static org.quartz.JobBuilder.newJob;
import static org.quartz.SimpleScheduleBuilder.repeatSecondlyForever;
import static org.quartz.TriggerBuilder.newTrigger; public class SchedulerMain { final static Logger logger = LoggerFactory.getLogger(SchedulerMain.class); public static void main(String[] args) throws Exception {
Scheduler scheduler = StdSchedulerFactory.getDefaultScheduler(); scheduler.start(); JobDetail jobDetail = newJob(HelloJob.class).build(); Trigger trigger = newTrigger()
.startNow()
.withSchedule(repeatSecondlyForever(2))
.build(); scheduler.scheduleJob(jobDetail, trigger);
} public static class HelloJob implements Job { @Override
public void execute(JobExecutionContext jobExecutionContext) throws JobExecutionException {
logger.info("HelloJob executed");
}
}
}

This simple example creates a HelloJob every two seconds that simply logs a message. Quartz has a very extensive API for creating Triggerschedules.

To test this application locally you can run the Maven build and then run the SchedulerMain Java class:

$ mvn package
INFO] Scanning for projects...
[INFO]
[INFO] ------------------------------------------------------------------------
[INFO] Building devcenter-java-quartz-rabbitmq 1.0-SNAPSHOT
[INFO] ------------------------------------------------------------------------ $ java -cp target/classes:target/dependency/* com.heroku.devcenter.SchedulerMain
...
66 [main] INFO org.quartz.impl.StdSchedulerFactory - Quartz scheduler 'DefaultQuartzScheduler' initialized from default resource file in Quartz package: 'quartz.properties'
66 [main] INFO org.quartz.impl.StdSchedulerFactory - Quartz scheduler version: 2.1.5
66 [main] INFO org.quartz.core.QuartzScheduler - Scheduler DefaultQuartzScheduler_$_NON_CLUSTERED started.
104 [DefaultQuartzScheduler_Worker-1] INFO com.heroku.devcenter.SchedulerMain - HelloJob executed
2084 [DefaultQuartzScheduler_Worker-2] INFO com.heroku.devcenter.SchedulerMain - HelloJob executed

Press Ctrl-C to exit the app.

If the HelloJob actually did work itself then we would have a runtime bottleneck because we could not scale the scheduler while avoiding duplicate jobs being scheduled. Quartz does have a JDBC module that can use a database to prevent jobs from being duplicated but a simpler approach is to only run one instance of the scheduler and have the scheduled jobs added to a message queue where they can be processes in parallel by job worker processes.

Queuing jobs with RabbitMQ

RabbitMQ can be used as a message queue so the scheduler process can be used just to add jobs to a queue and worker processes can be used to grab jobs from the queue and process them. To add the RabbitMQ client library as a dependency in Maven specify the following in dependencies block of the pom.xml file:

<dependency>
<groupId>com.rabbitmq</groupId>
<artifactId>amqp-client</artifactId>
<version>2.8.2</version>
</dependency>

If you want to test this locally then install RabbitMQ and set an environment variable that will provide the application the connection information to your RabbitMQ server.

On Windows:

$ set CLOUDAMQP_URL="amqp://guest:guest@localhost:5672/%2f"

On Mac/Linux:

$ export CLOUDAMQP_URL="amqp://guest:guest@localhost:5672/%2f"

The CLOUDAMQP_URL environment variable will be used by the scheduler and worker processes to connect to the shared message queue. This example uses that environment variable because that is the way theCloudAMQP Heroku Add-on will provide it’s connection information to the application.

The SchedulerMain class needs to be updated to add a new message onto a queue every time the HelloJob is executed. Here are the newSchedulerMain and HelloJob classes from the SchedulerMain.java file in the sample project:

package com.heroku.devcenter;

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;
import com.rabbitmq.client.MessageProperties;
import org.quartz.*;
import org.quartz.impl.StdSchedulerFactory;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory; import java.util.HashMap;
import java.util.Map; import static org.quartz.JobBuilder.newJob;
import static org.quartz.SimpleScheduleBuilder.repeatSecondlyForever;
import static org.quartz.TriggerBuilder.newTrigger; public class SchedulerMain { final static Logger logger = LoggerFactory.getLogger(SchedulerMain.class);
final static ConnectionFactory factory = new ConnectionFactory(); public static void main(String[] args) throws Exception {
factory.setUri(System.getenv("CLOUDAMQP_URL"));
Scheduler scheduler = StdSchedulerFactory.getDefaultScheduler(); scheduler.start(); JobDetail jobDetail = newJob(HelloJob.class).build(); Trigger trigger = newTrigger()
.startNow()
.withSchedule(repeatSecondlyForever(5))
.build(); scheduler.scheduleJob(jobDetail, trigger);
} public static class HelloJob implements Job { @Override
public void execute(JobExecutionContext jobExecutionContext) throws JobExecutionException { try {
Connection connection = factory.newConnection();
Channel channel = connection.createChannel();
String queueName = "work-queue-1";
Map<String, Object> params = new HashMap<String, Object>();
params.put("x-ha-policy", "all");
channel.queueDeclare(queueName, true, false, false, params); String msg = "Sent at:" + System.currentTimeMillis();
byte[] body = msg.getBytes("UTF-8");
channel.basicPublish("", queueName, MessageProperties.PERSISTENT_TEXT_PLAIN, body);
logger.info("Message Sent: " + msg);
connection.close();
}
catch (Exception e) {
logger.error(e.getMessage(), e);
} } } }

In this example every time the HelloJob is executed it adds a message onto a RabbitMQ message queue simply containing a String with the time the String was created. Running the updated SchedulerMainshould add a new message to the queue every 5 seconds.

Processing jobs

Next, create a Java application that will pull messages from the queue and handle them. This application will also use the RabbitFactoryUtilto get a connection to RabbitMQ from the CLOUDAMQP_URL environment variable. Here is the WorkerMain class from the WorkerMain.java file in the example project:

package com.heroku.devcenter;

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;
import com.rabbitmq.client.QueueingConsumer;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory; import java.util.Collections;
import java.util.HashMap;
import java.util.Map; public class WorkerMain { final static Logger logger = LoggerFactory.getLogger(WorkerMain.class); public static void main(String[] args) throws Exception { ConnectionFactory factory = new ConnectionFactory();
factory.setUri(System.getenv("CLOUDAMQP_URL"));
Connection connection = factory.newConnection();
Channel channel = connection.createChannel();
String queueName = "work-queue-1";
Map<String, Object> params = new HashMap<String, Object>();
params.put("x-ha-policy", "all");
channel.queueDeclare(queueName, true, false, false, params);
QueueingConsumer consumer = new QueueingConsumer(channel);
channel.basicConsume(queueName, false, consumer); while (true) {
QueueingConsumer.Delivery delivery = consumer.nextDelivery();
if (delivery != null) {
String msg = new String(delivery.getBody(), "UTF-8");
logger.info("Message Received: " + msg);
channel.basicAck(delivery.getEnvelope().getDeliveryTag(), false);
}
} } }

This class simply waits for new messages on the message queue and logs that it received them. You can run this example locally by doing a build and then running the WorkerMain class:

$ mvn package
$ java -cp target/classes:target/dependency/* com.heroku.devcenter.WorkerMain

You can also run multiple instances of this example locally to see how the job processing can be horizontally distributed.

Running on Heroku

Now that you have everything working locally you can run this on Heroku. First declare the process model in a new file named Procfilecontaining:

scheduler: java $JAVA_OPTS -cp target/classes:target/dependency/* com.heroku.devcenter.SchedulerMain
worker: java $JAVA_OPTS -cp target/classes:target/dependency/* com.heroku.devcenter.WorkerMain

This defines two process types that can be executed on Heroku; one named scheduler for the SchedulerMain app and one named workerfor the WorkerMain app.

To run on Heroku you will need to push a Git repository to Heroku containing the Maven build descriptor, source code, and Procfile. If you cloned the example project then you already have a Git repository. If you need to create a new git repository containing these files, run:

$ git init
$ git add src pom.xml Procfile
$ git commit -m init

Create a new application on Heroku from within the project’s root directory:

$ heroku create
Creating furious-cloud-2945... done, stack is cedar-14
http://furious-cloud-2945.herokuapp.com/ | git@heroku.com:furious-cloud-2945.git
Git remote heroku added

Then add the CloudAMQP add-on to your application:

$ heroku addons:add cloudamqp
Adding cloudamqp to furious-cloud-2945... done, v2 (free)
cloudamqp documentation available at: https://devcenter.heroku.com/articles/cloudamqp

Now push your Git repository to Heroku:

$ git push heroku master
Counting objects: 165, done.
Delta compression using up to 2 threads.
...
-----> Heroku receiving push
-----> Java app detected
...
-----> Discovering process types
Procfile declares types -> scheduler, worker
-----> Compiled slug size is 1.4MB
-----> Launching... done, v5
http://furious-cloud-2945.herokuapp.com deployed to Heroku

This will run the Maven build for your project on Heroku and create a slug file containing the executable assets for your application. To run the application you will need to allocate dynos to run each process type. You should only allocate one dyno to run the scheduler process type to avoid duplicate job scheduling. You can allocate as many dynos as needed to the worker process type since it is event driven and parallelizable through the RabbitMQ message queue.

To allocate one dyno to the scheduler process type run:

$ heroku ps:scale scheduler=1
Scaling scheduler processes... done, now running 1

This should begin adding messages to the queue every five seconds. To allocate two dynos to the worker process type run:

$ heroku ps:scale worker=2
Scaling worker processes... done, now running 2

This will provision two dynos, each which will run the WorkerMain app and pull messages from the queue for processing. You can verify that this is happening by watching the Heroku logs for your application. To open a feed of your logs run:

$ heroku logs -t
2012-06-26T22:26:47+00:00 app[scheduler.1]: 100223 [DefaultQuartzScheduler_Worker-1] INFO com.heroku.devcenter.SchedulerMain - Message Sent: Sent at:1340749607126
2012-06-26T22:26:47+00:00 app[worker.2]: 104798 [main] INFO com.heroku.devcenter.WorkerMain - Message Received: Sent at:1340749607126
2012-06-26T22:26:52+00:00 app[scheduler.1]: 105252 [DefaultQuartzScheduler_Worker-2] INFO com.heroku.devcenter.SchedulerMain - Message Sent: Sent at:1340749612155
2012-06-26T22:26:52+00:00 app[worker.1]: 109738 [main] INFO com.heroku.devcenter.WorkerMain - Message Received: Sent at:1340749612155

In this example execution the scheduler creates 2 messages which are handled by the two different worker dynos (worker.1 and worker.2). This shows that the work is being scheduled and distributed correctly.

Further learning

This example application just shows the basics for architecting a scalable and reliable system for scheduling and processing background jobs. To learn more see:

Scheduled Jobs with Custom Clock Processes in Java with Quartz and RabbitMQ的更多相关文章

  1. java 多线程——quartz 定时调度的例子

    java 多线程 目录: Java 多线程——基础知识 Java 多线程 —— synchronized关键字 java 多线程——一个定时调度的例子 java 多线程——quartz 定时调度的例子 ...

  2. 【RabbitMQ】 Java简单的实现RabbitMQ

    准备工作 1.安装RabbitMQ,参考[RabbitMQ] RabbitMQ安装 2.新建Java项目,引入RabbitMQ的Maven依赖 <dependency> <group ...

  3. java框架---->quartz的使用(一)

    Quartz 是个开源的作业调度框架,为在 Java 应用程序中进行作业调度提供了简单却强大的机制.今天我们就来学习一下它的使用,这里会分篇章对它进行介绍.只是希望能有个人,在我说没事的时候,知道我不 ...

  4. [Spring] Java spring quartz 定时任务

    首先,需要导入quartz 的jar包 ① applicationContext.xml <!-- 轮询任务 --> <import resource="classpath ...

  5. Java学习---Quartz定时任务快速入门

    Quartz是OpenSymphony开源组织在Job scheduling领域又一个开源项目,它可以与J2EE与J2SE应用程序相结合也可以单独使用.Quartz可以用来创建简单或为运行十个,百个, ...

  6. java任务调度quartz框架的小例子

    quartz是一个开源的作业调度框架,当然,java可以使用Timer来实现简单任务调度的功能,但Timer是单线程的设计方案,使得一个任务延迟会影响到其他的任务.java也可以使用Scheduled ...

  7. rabbitMQ第二篇:java简单的实现RabbitMQ

    前言:在这里我将用java来简单的实现rabbitMQ.下面我们带着下面问题来一步步的了解和学习rabbitMQ. 1:如果消费者连接中断,这期间我们应该怎么办 2:如何做到负载均衡 3:如何有效的将 ...

  8. java maven quartz exampe 实用指南

    pom.xml <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w ...

  9. java实现Quartz定时功能

    本文是在学习中的总结,欢迎转载但请注明出处:http://blog.csdn.net/pistolove/article/details/49975443 最近在学习定时相关的技术.当前,几乎所有的互 ...

随机推荐

  1. Struts 2 Tutorial Basic MVC Architecture

    Model View Controller or MVC as it is popularly called, is a software design pattern for developing ...

  2. Delphi与Socket

    一.Delphi与Socket计算机网络是由一系列网络通信协议组成的,其中的核心协议是传输层的TCPIP和UDP协议.TCP是面向连接的,通信双方保持一条通路,好比目前的电话线,使用telnet登陆B ...

  3. C#设置窗体中的窗体随主窗体大小变化而变化

    form2 f=new form2(); f.Size=this.Size; f.Location=this.Location; f.showdialog(); 作者:耑新新,发布于  博客园 转载请 ...

  4. 《精通Python设计模式》学习之工厂方法

    小书,在我以前作数据库的连接时,就用了这个工厂方法的. 归纳总结一下,更有利于成长吧. import xml.etree.ElementTree as etree import json class ...

  5. android拾遗——Android之Service与IntentService的比较

    不知道大家有没有和我一样,以前做项目或者练习的时候一直都是用Service来处理后台耗时操作,却很少注意到还有个IntentService,前段时间准备面试的时候看到了一篇关于IntentServic ...

  6. markdown转换为html

    想要自己实现markdown编辑器,欲使用markdown-it作为编辑器,有着比较多的插件,可以实现代码高亮以及对数学公式转换等功能. // Activate/deactivate rules, w ...

  7. windows7无声音,提示未插入扬声器或耳机的解决

    windows7无声音,提示未插入扬声器或耳机的解决: http://jingyan.baidu.com/article/358570f6043a85ce4624fc47.html

  8. 在centOS使用systemctl配置启动多个tomcat

    公司服务器使用的是阿里云CentOS7,CentOS7和CentOS6目前最大区别就是service变成了现在的systemctl,简单的查了一下并结合使用,发现systemctl功能上等同于6上面的 ...

  9. [leetcode greedy]455. Assign Cookies

    Assume you are an awesome parent and want to give your children some cookies. But, you should give e ...

  10. 【HDU 3662】3D Convex Hull

    http://acm.hdu.edu.cn/showproblem.php?pid=3662 求给定空间中的点的三维凸包上有多少个面. 用增量法,不断加入点,把新加的点能看到的面都删掉,不能看到的面与 ...