Scheduled Jobs with Custom Clock Processes in Java with Quartz and RabbitMQ
原文地址: https://devcenter.heroku.com/articles/scheduled-jobs-custom-clock-processes-java-quartz-rabbitmq
Table of Contents
- Prerequisites
- Scheduling jobs with Quartz
- Queuing jobs with RabbitMQ
- Processing jobs
- Running on Heroku
- Further learning
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
- The Heroku Toolbelt installed.
- Maven 3.0.4 installed.
- A Heroku user account. Signup is free and instant.
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
A 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
- 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 ...
- 【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 ...
- 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 ...
- [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 ...
- 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, ...
- 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 ...
- 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 ...
- 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 ...
- 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
- 微信小程序蓝牙模块
蓝牙部分知识 关于Service: 每个设备包含有多个Service,每个Service对应一个uuid 关于Characteristic 每个Service包含多个Characteristic,每个 ...
- 20165333 2016-2017-2 《Java程序设计》第1周学习总结
20165333 2016-2017-2 <Java程序设计>第1周学习总结 教材学习内容总结 java 的地位 Java 的特点 安装JDK 系统环境的设置 Java程序的编写,编译和运 ...
- Mongodb配置:error:10061 由于目标计算机积极拒绝,无法连接
相信很多学Node的同学,在进入MongoDB后台管理 Shell的时候都会“遇到error:10061 由于目标计算机积极拒绝,无法连接”这种情况,很多情况都是dbpath与dblog的路径没有配置 ...
- 【LOJ】#2289. 「THUWC 2017」在美妙的数学王国中畅游
题解 我们发现,题目告诉我们这个东西就是一个lct 首先,如果只有3,问题就非常简单了,我们算出所有a的总和,所有b的总和就好了 要是1和2也是多项式就好了--其实可以!也就是下面泰勒展开的用处,我们 ...
- USACO 6.5 Closed Fences
Closed Fences A closed fence in the plane is a set of non-crossing, connected line segments with N c ...
- 005 Ajax中使用jquery实现三种格式的信息
1.jquery中的ajax 二:load 2.load方法 3.load测试程序大纲 4.load测试程序 <!DOCTYPE html> <html> <head&g ...
- Python并发编程系列之常用概念剖析:并行 串行 并发 同步 异步 阻塞 非阻塞 进程 线程 协程
1 引言 并发.并行.串行.同步.异步.阻塞.非阻塞.进程.线程.协程是并发编程中的常见概念,相似却也有却不尽相同,令人头痛,这一篇博文中我们来区分一下这些概念. 2 并发与并行 在解释并发与并行之前 ...
- NetStandard类库实现Log4Net集成
前面都是Log4Net集成到NetCore项目中,集成到NetStandard类库还是第一次,所以记录一下 小提示:NetStandard要想同时被NetCore和NetFramework调用,需要在 ...
- Redis 服务器命令
1.BGREWRITEAOF 异步执行一个 AOF(AppendOnly File) 文件重写操作 2.BGSAVE 在后台异步保存当前数据库的数据到磁盘 3.CLIENT KILL [ip:port ...
- hihocoder 1509 异或排序
题面在这里! 考虑前后两个数 x,y,可以发现S只有在(x xor y)的最高有1位上的取值是要被确定的 (如果x==y那么没有限制),可以推一下什么情况下是1/0. 于是我们模拟一下这个操作,判一判 ...