分布式调度QUARTZ+SPRING

2019独角兽企业重金招聘Python工程师标准>>>

使用SPRING的定时任务框架，如果是在分布式的环境下，由于有多台节点，会产生相同的任务，会被多个节点执行，这时需引入分布式的QUARTZ。
触发器：存放时间排程
任务：蔟业务代码
排程器：负责调度，即在指定的时间执行对应的任务

如果是分布式QUARTZ，则各个节点会上报任务，存到数据库中，执行时会从数据库中取出触发器来执行，如果触发器的名称和执行时间相同，则只有一个节点去执行此任务。
如果此节点执行失败，则此任务则会被分派到另一节点执行。

quartz.properties

# ============================================================================
#  Configure JobStore  
#  Using Spring datasource in quartzJobsConfig.xml
#  Spring uses LocalDataSourceJobStore extension of JobStoreCMT
# ============================================================================
org.quartz.jobStore.useProperties=true
org.quartz.jobStore.tablePrefix = QRTZ_
org.quartz.jobStore.isClustered = true
org.quartz.jobStore.clusterCheckinInterval = 5000
org.quartz.jobStore.misfireThreshold = 60000
org.quartz.jobStore.txIsolationLevelReadCommitted = true
 
#  Change this to match your DB vendor
org.quartz.jobStore. class = org.quartz.impl.jdbcjobstore.JobStoreTX
org.quartz.jobStore.driverDelegateClass = org.quartz.impl.jdbcjobstore.StdJDBCDelegate
 

# ============================================================================
#  Configure Main Scheduler Properties  
#  Needed to manage cluster instances
# ============================================================================
org.quartz.scheduler.instanceId=AUTO
org.quartz.scheduler.instanceName=MY_CLUSTERED_JOB_SCHEDULER
org.quartz.scheduler.rmi.export = false
org.quartz.scheduler.rmi.proxy = false


# ============================================================================
#  Configure ThreadPool  
# ============================================================================
org.quartz.threadPool. class = org.quartz.simpl.SimpleThreadPool
org.quartz.threadPool.threadCount = 10
org.quartz.threadPool.threadPriority = 5
org.quartz.threadPool.threadsInheritContextClassLoaderOfInitializingThread = true

web-schedule-applicationcontext.xml

<? xml version="1.0" encoding="UTF-8" ?>
< beans  xmlns ="http://www.springframework.org/schema/beans"
    xmlns:xsi ="http://www.w3.org/2001/XMLSchema-instance"  xmlns:context ="http://www.springframework.org/schema/context"
    xmlns:mongo ="http://www.springframework.org/schema/data/mongo"
    xsi:schemaLocation ="http://www.springframework.org/schema/context
          http://www.springframework.org/schema/context/spring-context-3.0.xsd
          http://www.springframework.org/schema/data/mongo
          http://www.springframework.org/schema/data/mongo/spring-mongo-1.3.xsd
          http://www.springframework.org/schema/beans
          http://www.springframework.org/schema/beans/spring-beans-3.0.xsd" >


     <!--  增加定时器配置  -->
     <!--  线程执行器配置，用于任务注册  -->
     < bean  id ="executor"  class ="org.springframework.scheduling.concurrent.ThreadPoolTaskExecutor" >
          < property  name ="corePoolSize"  value ="10"   />
          < property  name ="maxPoolSize"  value ="100"   />
          < property  name ="queueCapacity"  value ="500"   />
     </ bean >

     <!--  设置调度  -->
     < bean  id ="webScheduler"
        class ="org.springframework.scheduling.quartz.SchedulerFactoryBean" >

         < property  name ="configLocation"  value ="classpath:/properties/config/quartz.properties"   />
         < property  name ="dataSource"  ref ="dataSourceCMS"   />
         < property  name ="transactionManager"  ref ="txManager"   />

         <!--  This name is persisted as SCHED_NAME in db. for local testing could 
            change to unique name to avoid collision with dev server  -->
         < property  name ="schedulerName"  value ="quartzScheduler"   />

         <!--  Will update database cron triggers to what is in this jobs file on 
            each deploy. Replaces all previous trigger and job data that was in the database. 
            YMMV  -->
         < property  name ="overwriteExistingJobs"  value ="true"   />

         < property  name ="startupDelay"  value ="5" />
         < property  name ="applicationContextSchedulerContextKey"  value ="applicationContext"   />
         < property  name ="jobFactory" >
             < bean  class ="com.tcl.project7.boss.common.scheduling.AutowiringSpringBeanJobFactory"   />
         </ property >
        
         < property  name ="triggers" >
               < list >
                        < ref  bean ="springQuertzClusterTaskSchedulerTesterTigger"   />
               </ list >
          </ property >
         < property  name ="jobDetails" >
             < list >
                 < ref  bean ="springQuertzClusterTaskSchedulerTesterJobDetail"   />
             </ list >
         </ property >
          < property  name ="taskExecutor"  ref ="executor"   />

     </ bean >


    
    
    
     <!--  触发器  -->
     < bean  id ="springQuertzClusterTaskSchedulerTesterTigger"  class ="common.scheduling.PersistableCronTriggerFactoryBean" >
         < property  name ="jobDetail"  ref ="springQuertzClusterTaskSchedulerTesterJobDetail" />
         < property  name ="cronExpression"  value ="* * * * * ?"   />    
     </ bean >
    
     < bean  id ="springQuertzClusterTaskSchedulerTesterJobDetail"  class ="org.springframework.scheduling.quartz.JobDetailBean" >
         < property  name ="jobClass"  value ="common.scheduling.SpringQuertzClusterTaskSchedulerTester"   />
        
         <!--  fail-over 重写执行失败的任务,default=false  -->
         < property  name ="requestsRecovery"  value ="false" />
     </ bean >
    
    
    
</ beans >

JOB文件：SpringQuertzClusterTaskSchedulerTester.java

package common.scheduling;

import java.util.Date;

import org.quartz.JobExecutionContext;
import org.quartz.JobExecutionException;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.scheduling.quartz.QuartzJobBean;

import com.tcl.project7.boss.common.util.UrlUtil;
import com.tcl.project7.boss.common.util.time.TimeUtils;
/**
 * <p>Title:SpringQuertzClusterTaskSchedulerTester</p>
 * <p>Description:
 * 应为要持久化等特性操作,需要继承 QuartzJobBean
 * <br>由于要被持久化,所以不能存放xxxxManager类似对象,
 * 只能从每次从QuartzJobBean注入的ApplicationContext 中去取出
 *
 * </p>    
 *
 *
  */
public  class SpringQuertzClusterTaskSchedulerTester  extends QuartzJobBean {
    
     private  static Logger logger = LoggerFactory.getLogger(SpringQuertzClusterTaskSchedulerTester. class);
    
    @Autowired
     private UrlUtil urlUtil;
    
    
     protected  void executeInternal(JobExecutionContext arg0)
             throws JobExecutionException {
        logger.info("------" + TimeUtils.formatTime( new Date()) + "------" + urlUtil.getNginxHost());
        System.out.println("------" + TimeUtils.formatTime( new Date()) + "------" + urlUtil.getNginxHost());
    }
    
}

如果JOB中有需要调用SPRING的BEAN，则需要此文件AutowiringSpringBeanJobFactory.java

package common.scheduling;

import org.quartz.spi.TriggerFiredBundle;
import org.springframework.beans.factory.config.AutowireCapableBeanFactory;
import org.springframework.context.ApplicationContext;
import org.springframework.context.ApplicationContextAware;
import org.springframework.scheduling.quartz.SpringBeanJobFactory;

/**
 * Autowire Quartz Jobs with Spring context dependencies
 *  @see   http://stackoverflow.com/questions/6990767/inject-bean-reference-into-a-quartz-job-in-spring/15211030 #15211030
  */
public  final  class AutowiringSpringBeanJobFactory  extends SpringBeanJobFactory  implements ApplicationContextAware {
    
     private  transient AutowireCapableBeanFactory beanFactory;
 
     public  void setApplicationContext( final ApplicationContext context) {
        beanFactory = context.getAutowireCapableBeanFactory();
    }
 
    @Override
     protected Object createJobInstance( final TriggerFiredBundle bundle)  throws Exception {
         final Object job =  super.createJobInstance(bundle);
        beanFactory.autowireBean(job);
         return job;
    }
}

由于JOB需要存储到数据库中，会产生PROPERTY的问题，需剔除JOB-DATA，需此文件PersistableCronTriggerFactoryBean.java

package common.scheduling;

import org.springframework.scheduling.quartz.CronTriggerFactoryBean;
import org.springframework.scheduling.quartz.JobDetailAwareTrigger;

/**
 * Needed to set Quartz useProperties=true when using Spring classes,
 * because Spring sets an object reference on JobDataMap that is not a String
 * 
 *  @see   http://site.trimplement.com/using-spring-and-quartz-with-jobstore-properties/
 *  @see   http://forum.springsource.org/showthread.php?130984-Quartz-error-IOException
  */
public  class PersistableCronTriggerFactoryBean  extends CronTriggerFactoryBean {
    @Override
     public  void afterPropertiesSet() {
         super.afterPropertiesSet();
 
         // Remove the JobDetail element
        getJobDataMap().remove(JobDetailAwareTrigger.JOB_DETAIL_KEY);
    }
}

建表语句，MYSQL：quartzTables.sql

#
# Quartz seems  to  work best  with the driver mm.mysql - 2.0. 7 -bin.jar
#
#  In your Quartz properties  file, you ' ll need to set 
# org.quartz.jobStore.driverDelegateClass = org.quartz.impl.jdbcjobstore.StdJDBCDelegate
#

DROP TABLE IF EXISTS QRTZ_JOB_LISTENERS;
DROP TABLE IF EXISTS QRTZ_TRIGGER_LISTENERS;
DROP TABLE IF EXISTS QRTZ_FIRED_TRIGGERS;
DROP TABLE IF EXISTS QRTZ_PAUSED_TRIGGER_GRPS;
DROP TABLE IF EXISTS QRTZ_SCHEDULER_STATE;
DROP TABLE IF EXISTS QRTZ_LOCKS;
DROP TABLE IF EXISTS QRTZ_SIMPLE_TRIGGERS;
DROP TABLE IF EXISTS QRTZ_CRON_TRIGGERS;
DROP TABLE IF EXISTS QRTZ_BLOB_TRIGGERS;
DROP TABLE IF EXISTS QRTZ_TRIGGERS;
DROP TABLE IF EXISTS QRTZ_JOB_DETAILS;
DROP TABLE IF EXISTS QRTZ_CALENDARS;


CREATE TABLE QRTZ_JOB_DETAILS
  (
    JOB_NAME  VARCHAR(200) NOT NULL,
    JOB_GROUP VARCHAR(200) NOT NULL,
    DESCRIPTION VARCHAR(250) NULL,
    JOB_CLASS_NAME   VARCHAR(250) NOT NULL,
    IS_DURABLE VARCHAR(1) NOT NULL,
    IS_VOLATILE VARCHAR(1) NOT NULL,
    IS_STATEFUL VARCHAR(1) NOT NULL,
    REQUESTS_RECOVERY VARCHAR(1) NOT NULL,
    JOB_DATA BLOB NULL,
    PRIMARY KEY (JOB_NAME,JOB_GROUP)
);

CREATE TABLE QRTZ_JOB_LISTENERS
  (
    JOB_NAME  VARCHAR(200) NOT NULL,
    JOB_GROUP VARCHAR(200) NOT NULL,
    JOB_LISTENER VARCHAR(200) NOT NULL,
    PRIMARY KEY (JOB_NAME,JOB_GROUP,JOB_LISTENER),
    FOREIGN KEY (JOB_NAME,JOB_GROUP)
        REFERENCES QRTZ_JOB_DETAILS(JOB_NAME,JOB_GROUP)
);

CREATE TABLE QRTZ_TRIGGERS
  (
    TRIGGER_NAME VARCHAR(200) NOT NULL,
    TRIGGER_GROUP VARCHAR(200) NOT NULL,
    JOB_NAME  VARCHAR(200) NOT NULL,
    JOB_GROUP VARCHAR(200) NOT NULL,
    IS_VOLATILE VARCHAR(1) NOT NULL,
    DESCRIPTION VARCHAR(250) NULL,
    NEXT_FIRE_TIME BIGINT(13) NULL,
    PREV_FIRE_TIME BIGINT(13) NULL,
    PRIORITY INTEGER NULL,
    TRIGGER_STATE VARCHAR(16) NOT NULL,
    TRIGGER_TYPE VARCHAR(8) NOT NULL,
    START_TIME BIGINT(13) NOT NULL,
    END_TIME BIGINT(13) NULL,
    CALENDAR_NAME VARCHAR(200) NULL,
    MISFIRE_INSTR SMALLINT(2) NULL,
    JOB_DATA BLOB NULL,
    PRIMARY KEY (TRIGGER_NAME,TRIGGER_GROUP),
    FOREIGN KEY (JOB_NAME,JOB_GROUP)
        REFERENCES QRTZ_JOB_DETAILS(JOB_NAME,JOB_GROUP)
);

CREATE TABLE QRTZ_SIMPLE_TRIGGERS
  (
    TRIGGER_NAME VARCHAR(200) NOT NULL,
    TRIGGER_GROUP VARCHAR(200) NOT NULL,
    REPEAT_COUNT BIGINT(7) NOT NULL,
    REPEAT_INTERVAL BIGINT(12) NOT NULL,
    TIMES_TRIGGERED BIGINT(10) NOT NULL,
    PRIMARY KEY (TRIGGER_NAME,TRIGGER_GROUP),
    FOREIGN KEY (TRIGGER_NAME,TRIGGER_GROUP)
        REFERENCES QRTZ_TRIGGERS(TRIGGER_NAME,TRIGGER_GROUP)
);

CREATE TABLE QRTZ_CRON_TRIGGERS
  (
    TRIGGER_NAME VARCHAR(200) NOT NULL,
    TRIGGER_GROUP VARCHAR(200) NOT NULL,
    CRON_EXPRESSION VARCHAR(200) NOT NULL,
    TIME_ZONE_ID VARCHAR(80),
    PRIMARY KEY (TRIGGER_NAME,TRIGGER_GROUP),
    FOREIGN KEY (TRIGGER_NAME,TRIGGER_GROUP)
        REFERENCES QRTZ_TRIGGERS(TRIGGER_NAME,TRIGGER_GROUP)
);

CREATE TABLE QRTZ_BLOB_TRIGGERS
  (
    TRIGGER_NAME VARCHAR(200) NOT NULL,
    TRIGGER_GROUP VARCHAR(200) NOT NULL,
    BLOB_DATA BLOB NULL,
    PRIMARY KEY (TRIGGER_NAME,TRIGGER_GROUP),
    FOREIGN KEY (TRIGGER_NAME,TRIGGER_GROUP)
        REFERENCES QRTZ_TRIGGERS(TRIGGER_NAME,TRIGGER_GROUP)
);

CREATE TABLE QRTZ_TRIGGER_LISTENERS
  (
    TRIGGER_NAME  VARCHAR(200) NOT NULL,
    TRIGGER_GROUP VARCHAR(200) NOT NULL,
    TRIGGER_LISTENER VARCHAR(200) NOT NULL,
    PRIMARY KEY (TRIGGER_NAME,TRIGGER_GROUP,TRIGGER_LISTENER),
    FOREIGN KEY (TRIGGER_NAME,TRIGGER_GROUP)
        REFERENCES QRTZ_TRIGGERS(TRIGGER_NAME,TRIGGER_GROUP)
);


CREATE TABLE QRTZ_CALENDARS
  (
    CALENDAR_NAME  VARCHAR(200) NOT NULL,
    CALENDAR BLOB NOT NULL,
    PRIMARY KEY (CALENDAR_NAME)
);



CREATE TABLE QRTZ_PAUSED_TRIGGER_GRPS
  (
    TRIGGER_GROUP  VARCHAR(200) NOT NULL, 
    PRIMARY KEY (TRIGGER_GROUP)
);

CREATE TABLE QRTZ_FIRED_TRIGGERS
  (
    ENTRY_ID VARCHAR(95) NOT NULL,
    TRIGGER_NAME VARCHAR(200) NOT NULL,
    TRIGGER_GROUP VARCHAR(200) NOT NULL,
    IS_VOLATILE VARCHAR(1) NOT NULL,
    INSTANCE_NAME VARCHAR(200) NOT NULL,
    FIRED_TIME BIGINT(13) NOT NULL,
    PRIORITY INTEGER NOT NULL,
    STATE VARCHAR(16) NOT NULL,
    JOB_NAME VARCHAR(200) NULL,
    JOB_GROUP VARCHAR(200) NULL,
    IS_STATEFUL VARCHAR(1) NULL,
    REQUESTS_RECOVERY VARCHAR(1) NULL,
    PRIMARY KEY (ENTRY_ID)
);

CREATE TABLE QRTZ_SCHEDULER_STATE
  (
    INSTANCE_NAME VARCHAR(200) NOT NULL,
    LAST_CHECKIN_TIME BIGINT(13) NOT NULL,
    CHECKIN_INTERVAL BIGINT(13) NOT NULL,
    PRIMARY KEY (INSTANCE_NAME)
);

CREATE TABLE QRTZ_LOCKS
  (
    LOCK_NAME  VARCHAR(40) NOT NULL, 
    PRIMARY KEY (LOCK_NAME)
);


INSERT INTO QRTZ_LOCKS values( 'TRIGGER_ACCESS ' );
INSERT INTO QRTZ_LOCKS values( 'JOB_ACCESS ' );
INSERT INTO QRTZ_LOCKS values( 'CALENDAR_ACCESS ' );
INSERT INTO QRTZ_LOCKS values( 'STATE_ACCESS ' );
INSERT INTO QRTZ_LOCKS values( 'MISFIRE_ACCESS ' );


commit;

转载于:https://my.oschina.net/pvpCC9IFwqz4/blog/512524