Question

I’m using Spring’s SchedulerFactoryBean to run some Quartz jobs within a Spring based java application. At present, this is a single instance application in development, but as soon as we start horizontally scaling this we will want to use a jdbc based JobStore for Quartz so no more than one app will run a given job.

Right now, SchedulerFactoryBean is configured as follows:

 <bean id="schedulerFactoryBean" class="org.springframework.scheduling.quartz.SchedulerFactoryBean" >       
     <property name="taskExecutor" ref="taskExecutor"/>
     <property name="triggers">
         <list>
             <!-- a bunch of triggers here -->
         </list>
     <property name="applicationContextSchedulerContextKey">
         <value>applicationContext</value>
     </property>
</bean>

and with using a jdbc based JobStore it will look like this

<bean id="schedulerFactoryBean" class="org.springframework.scheduling.quartz.SchedulerFactoryBean" >
        <property name="dataSource" ref="mysqlJobDataSource"/>
        <property name="taskExecutor" ref="taskExecutor"/>
        <property name="triggers">
            <list>
                <!-- a bunch of triggers here -->
            </list>
        </property>
        <property name="applicationContextSchedulerContextKey">
            <value>applicationContext</value>
        </property>
        <property name="quartzProperties">
            <props>
                <prop key="org.quartz.jobStore.class">org.quartz.impl.jdbcjobstore.JobStoreTX</prop>
                <prop key="org.quartz.jobStore.driverDelegateClass">org.quartz.impl.jdbcjobstore.StdJDBCDelegate</prop>
                <!-- and a bunch of other quartz props -->
            </props>
        </property>
    </bean>

Ideally, I’d like to continue using the default RAMJobStore version (the first one) for developers, but use the jdbc version for deployed environments. However, there doesn’t seem to be a very good way to switch between the two through something like a property, since the jdbc store involves lots more configuration and the mere existence of the dataSource property on SchedulerFactoryBean means it tries to a JDBC based job store.

Also, Since SchedulerFactoryBean is an initializing bean where the initializing basically starts running all of the jobs, so I can’t have both of those beans defined in a config file loaded into the spring context either, which means I’ll have parallel jobs running.

I’ve also read through this answer, but this situtation differs in that I’m dealing with two InitializingBeans that should never be in the same context at the same time.

What would be the simplest way to configure switching between these two configurations of SchedulerFactoryBean?

Answer 1

From Spring 3.1 you can use Spring profiles:

<bean name="schedulerFactoryBean" profile="dev" ...

<bean name="schedulerFactoryBean" profile="prd" ...

Then you can instruct Spring container which profile to use, see How to set active spring 3.1 environment profile via a properites file and not via an env variable or system property and Spring autowire a stubbed service – duplicate bean.

If you can’t use 3.1 or profiles, the old-school of solving such issues is to have two context files: schedulerContext-dev.xml and schedulerContext-prd.xml`. Then you can import them selectively:

<import resource="schedulerContext-${some.property}"/>