I have two independent spring batch jobs in the same project because I want to use the same infrastructure-related beans. Everything is configured in Java. I would like to know if there's a proper way to start the jobs independent based for example on the first java app argument in the main method for example. If I run SpringApplication.run only the second job gets executed by magic.
The main method looks like:
@ComponentScan
@EnableAutoConfiguration
public class Application {
public static void main(String[] args) {
SpringApplication app = new SpringApplication(Application.class);
app.setWebEnvironment(false);
ApplicationContext ctx= app.run(args);
}
}
and the two jobs are configured as presented in the Spring Batch Getting Started tutorial on Spring.io. Here is the configuration file of the first job, the second being configured in the same way.
@Configuration
@EnableBatchProcessing
@Import({StandaloneInfrastructureConfiguration.class, ServicesConfiguration.class})
public class AddPodcastJobConfiguration {
@Autowired
private JobBuilderFactory jobs;
@Autowired
private StepBuilderFactory stepBuilderFactory;
//reader, writer, processor...
}
To enable modularization I created an AppConfig class, where I define factories for the two jobs:
@Configuration
@EnableBatchProcessing(modular=true)
public class AppConfig {
@Bean
public ApplicationContextFactory addNewPodcastJobs(){
return new GenericApplicationContextFactory(AddPodcastJobConfiguration.class);
}
@Bean
public ApplicationContextFactory newEpisodesNotificationJobs(){
return new GenericApplicationContextFactory(NotifySubscribersJobConfiguration.class);
}
}
P.S. I am new to Spring configuration in Java configuration Spring Boot and Spring Batch...
To run the jobs you like from the main method you can load the the required job configuration bean and the JobLauncher from the application context and then run it:
@ComponentScan
@EnableAutoConfiguration
public class ApplicationWithJobLauncher {
public static void main(String[] args) throws BeansException, JobExecutionAlreadyRunningException, JobRestartException, JobInstanceAlreadyCompleteException, JobParametersInvalidException, InterruptedException {
Log log = LogFactory.getLog(ApplicationWithJobLauncher.class);
SpringApplication app = new SpringApplication(ApplicationWithJobLauncher.class);
app.setWebEnvironment(false);
ConfigurableApplicationContext ctx= app.run(args);
JobLauncher jobLauncher = ctx.getBean(JobLauncher.class);
JobParameters jobParameters = new JobParametersBuilder()
.addDate("date", new Date())
.toJobParameters();
if("1".equals(args[0])){
//addNewPodcastJob
Job addNewPodcastJob = ctx.getBean("addNewPodcastJob", Job.class);
JobExecution jobExecution = jobLauncher.run(addNewPodcastJob, jobParameters);
} else {
jobLauncher.run(ctx.getBean("newEpisodesNotificationJob", Job.class), jobParameters);
}
System.exit(0);
}
}
What was causing my lots of confusion was that the second job were executed, even though the first job seemed to be "picked up" by the runner... Well the problem was that in both job's configuration file I used standard method names writer(), reader(), processor() and step() and it used the ones from the second job that seemed to "overwrite" the ones from the first job without any warnings...
I used though an application config class with @EnableBatchProcessing(modular=true), that I thought would be used magically by Spring Boot :
@Configuration
@EnableBatchProcessing(modular=true)
public class AppConfig {
@Bean
public ApplicationContextFactory addNewPodcastJobs(){
return new GenericApplicationContextFactory(AddPodcastJobConfiguration.class);
}
@Bean
public ApplicationContextFactory newEpisodesNotificationJobs(){
return new GenericApplicationContextFactory(NotifySubscribersJobConfiguration.class);
}
}
I will write a blog post about it when it is ready, but until then the code is available at https://github.com/podcastpedia/podcastpedia-batch (work/learning in progress)..