Why is my Spring Batch Task launching with the same JOB_INSTANCE_ID for multiple job executions? - mysql

I have a Spring Batch Task running on our cloud platform that will launch with the provided command line parameters, and then skip over the execution of the first Step with the following error:
[OUT] The job execution id 992 was run within the task execution 1325
[OUT] Step already complete or not restartable, so no action to execute:
StepExecution: id=1071, version=3, name=OFileStep, status=COMPLETED, exitStatus=COMPLETED, readCount=0, filterCount=0, writeCount=0 readSkipCount=0,
writeSkipCount=0, processSkipCount=0, commitCount=1, rollbackCount=0, exitDescription=
I have investigated the metadata tables in the MySQL instance that Spring Batch uses to find that the JOB_INSTANCE_ID is the same between multiple executions, when it should increment by 1 each time.
The #Bean that I have defined for the Job Configuration is:
#Bean
public Job job() {
return jobBuilderFactory.get(OTaskConstants.JOB_NAME)
.listener(listener())
.incrementer(new RunIdIncrementer())
.start(dataTransferTaskStep())
.next(controlMTaskStep())
.build();
}
Is anyone aware of what could be causing this behavior?

Below line clearly says it all.
Step already complete or not restartable, so no action to execute:
Meaning the step/job already complete and can not be restarted. This is the behavior of Spring Batch. In order to by pass this we need to pass an unique argument.
In your case i see you already have RunIdIncrementer. Now question is why it is not working.
Can you see BATCH_JOB_PARMS table to see what arguments are getting passed to the job? May be you are missing something.
You can also use SimpleIncrementor. See below code for explanation.
https://docs.spring.io/spring-batch/docs/current/reference/html/index-single.html#JobParametersIncrementer

Remove #Bean annotation on Job.
It causes the Job to be launched with no parameters every time you launch/start application as spring tries to load the bean and which in-turn launches the batch job.
Remove the annotation and use spring scheduler to schedule the jobs.

I had the same issue. Below code helped me resolve it. By adding params in job launcher a new job_instance_id is created for every run.
#SpringBootApplication
public class App implements CommandLineRunner {
#Autowired
JobLauncher jobLauncher;
#Autowired
Job job;
public static void main(String[] args) {
SpringApplication.run(App.class, args);
}
#Override
public void run(String... args) throws Exception {
JobParameters params = new JobParametersBuilder()
.addString("JobID", String.valueOf(System.currentTimeMillis()))
.toJobParameters();
jobLauncher.run(job, params);
}
}

Solution
Refer error message above “If you want to run this job again, change the parameters.” The formula is JobInstance = JobParameters + Job. If you do not have any parameters for JobParameters, just pass a current time as parameter to create a new JobInstance. For example,
CustomJobLauncher.java
//...
#Component
public class CustomJobLauncher {
#Autowired
JobLauncher jobLauncher;
#Autowired
Job job;
public void run() {
try {
JobParameters jobParameters =
new JobParametersBuilder()
.addLong("time",System.currentTimeMillis()).toJobParameters();
JobExecution execution = jobLauncher.run(job, jobParameters);
System.out.println("Exit Status : " + execution.getStatus());
} catch (Exception e) {
e.printStackTrace();
}
}
}
Source : https://mkyong.com/spring-batch/spring-batch-a-job-instance-already-exists-and-is-complete-for-parameters/

Related

Spring-boot Redis JMS JUnit

I am using Redis Server for message broker in my spring boot application.
Is there any simple way to Junit my publish and receive API?
e.g :
Publisher :
public String publish(Object domainObj) {
template.convertAndSend(topic.getTopic(), domainObj.toString());
return "Event Published";
}
Receiver :
public class Receiver implements MessageListener {
#Override
public void onMessage(Message message, byte[] bytes) {
System.out.println("Consumed Message {}" + message);
}
}
I am using JedisConnectionFactory and RedisMessageListenerContainer and RedisTemplate for my implementation
#Configuration
#EnableRedisRepositories
public class RedisConfig {
#Bean
public JedisConnectionFactory connectionFactory() {
RedisStandaloneConfiguration configuration = new RedisStandaloneConfiguration();
configuration.setHostName("localhost");
configuration.setPort(6379);
return new JedisConnectionFactory(configuration);
}
#Bean
public RedisTemplate<String, Object> template() {
RedisTemplate<String, Object> template = new RedisTemplate<>();
template.setConnectionFactory(connectionFactory());
template.setKeySerializer(new StringRedisSerializer());
template.setHashKeySerializer(new StringRedisSerializer());
template.setHashKeySerializer(new JdkSerializationRedisSerializer());
template.setValueSerializer(new JdkSerializationRedisSerializer());
template.setEnableTransactionSupport(true);
template.afterPropertiesSet();
return template;
}
#Bean
public ChannelTopic topic() {
return new ChannelTopic("common-channel");
}
#Bean
public MessageListenerAdapter messageListenerAdapter() {
return new MessageListenerAdapter(new Receiver());
}
#Bean
public RedisMessageListenerContainer redisMessageListenerContainer() {
RedisMessageListenerContainer container = new RedisMessageListenerContainer();
container.setConnectionFactory(connectionFactory());
container.addMessageListener(messageListenerAdapter(), topic());
return container;
}
Unit Testing Receiver and Publisher implementation is quite straight.
JUnit 5 coupled with Mockito extension should do the job.
For example for testing that :
public String publish(Object domainObj) {
template.convertAndSend(topic.getTopic(), domainObj.toString());
return "Event Published";
}
I expect that topic and template be fields of the current class.
These fields could be set by constructor.
So you could write something that check that convertAndSend() is eventually executed with the correct parameters :
#Mock
RedisTemplate<String, Object> templateMock;
#Test
void publish(){
Topic topicFixture = new Topic(...);
Object domainObjFixture = new FooBar(...);
Publisher publisher = new Publisher(templateMock, topicFixture);
//when
publisher.publish(domainObjFixture);
// then
Mockito.verify(templateMock)
.convertAndSend(topicFixture.getTopic(), domainObjFixture);
}
But I don't think that the unit test of these two classes be enough because it never tests the final things : the JMS processing performed by Redis backend.
Particularly, the RedisConfig part that you set with specific things as serializers that have important side effects on the processing.
For my part, I try to always write integration or partial integration tests for Redis backend stuffs to ensure a good no regression harness.
The java embedded-redis library is good for that. It allows to start a redis server
on localhost (works on Windows as well as on Linux).
Starting and stopping the redis server is as simple as :
RedisServer redisServer = new RedisServer(6379);
redisServer.start();
// do some work
redisServer.stop();
Move the start() in the #BeforeEach and the stop() in the #AfterEach and the server is ready.
Then it still requires some adjustments to ensure that the redis configuration specified in Spring is well setup during the tests while using your local redis server and not the "real" redis server. Not always simple to set but great when it is done !
The simplest way to unit test this is to use embedded-redis module. What you do is in BeforeAll you can start embedded Redis and stop the embedded Redis in AfterAll method.
You can also PostConstruct PreDestroy annotations to accomplish this.
If you're looking for Junit5 then you can find the code in my repo here
See BootstrapRedis annotation and their usage here
https://github.com/sonus21/rqueue/blob/7ef545c15985ef91ba719f070f7cc80745525047/rqueue-core/src/test/java/com/github/sonus21/rqueue/core/RedisScriptFactoryTest.java#L40

ShedLock - Not Executing

I am using shedlock library 4.20.0.
net.javacrumbs.shedlock shedlock-spring 4.20.0 net.javacrumbs.shedlock shedlock-provider-jdbc-template 2.1.0
The scheduler job is,
#scheduled(fixedRate = 5000)
#SchedulerLock(name = "TaskScheduler__scheduledTask", lockAtLeastForString = "PT5M", lockAtMostForString = "PT14M")
public void reportCurrentTime() {
LockAssert.assertLocked();
log.info("The time is now {} {}", dateFormat.format(new Date()), dataSource);
}
It shows #SchedulerLock as deprecated.
And the spring boot class,
#SpringBootApplication
#EnableScheduling
#EnableSchedulerLock(defaultLockAtMostFor = "PT30S")
public class DMSCaseEmulatorSpringApplication {
public static void main(String[] args) {
SpringApplication.run(DMSCaseEmulatorSpringApplication.class, args);
}
}
When i execute the spring boot class, it triggers shedlock and creates a record in database table but in logs i keep getting as below,
19:54:39.188 [scheduling-1] DEBUG n.j.s.c.DefaultLockingTaskExecutor - Locked TaskScheduler__scheduledTask.
19:54:39.188 [scheduling-1] INFO u.g.h.c.d.s.ScheduledTasks - The time is now 19:54:39 HikariDataSource (HikariPool-1)
19:54:39.205 [scheduling-1] DEBUG n.j.s.c.DefaultLockingTaskExecutor - Unlocked TaskScheduler__scheduledTask.
19:54:44.065 [scheduling-1] DEBUG n.j.s.c.DefaultLockingTaskExecutor - Not executing TaskScheduler__scheduledTask. It's locked.
19:54:49.062 [scheduling-1] DEBUG n.j.s.c.DefaultLockingTaskExecutor - Not executing TaskScheduler__scheduledTask. It's locked.
Any thoughts will be appreciated?
The issue is caused by lockAtLeastForString = "PT5M" By specifying that, you are saying that the lock should be held at least for 5 minutes even if the task finishes sooner.
Regarding the Deprecation warning, please consult the JavaDoc.

Assertion error while running test with LatchCountDownAndCallRealMethodAnswer from RabbitMQ

I have a listener test, where i post a message in a parallel thread and check with LatchCountDownAndCallRealMethodAnswer if the all were processed successfully. Running the test alone, it works perfectly, however if you run all other tests together, it fails because it failed to leave the counter at zero, but the listener received and processed the message normally. Does anyone have any ideas?
My Test Class
#RunWith(SpringRunner.class)
#SpringBootTest
#RabbitListenerTest
#ActiveProfiles("test")
public class EventListenerTest {
EventListener eventListener;
#Autowired
protected RabbitListenerTestHarness harness;
#Autowired
private EventStoreRepository repository;
#SpyBean
private DomainEventPublisher publisher;
#Before
public void setUp() {
MockitoAnnotations.initMocks(this);
DomainRegister.setDomainEventPublisher(publisher);
eventListener = this.harness.getSpy("eventListenerId");
}
#Test
public void storeEventsListenerTest() throws Exception {
LatchCountDownAndCallRealMethodAnswer answer = new LatchCountDownAndCallRealMethodAnswer(1);
doAnswer(answer).when(eventListener).storeEvents(any(BalanceReserved.class));
publisher.publish(new BalanceReserved("12233", 150.0, BigDecimal.ZERO), "");
assertTrue(answer.getLatch().await(10, TimeUnit.SECONDS));
verify(eventListener, times(1)).storeEvents(any(BalanceReserved.class));
}
#After
public void tearDown() {
DomainRegister.setDomainEventPublisher(null);
reset(eventListener);
repository.deleteAll();
}
}
Error
java.lang.AssertionError
If you have other tests using the same queue, you need to shut down the application context for each test so the test's listeners are stopped. By default, the Spring Test framework caches the application context for reuse. This will cause other tests to "steal" messages.
Add #DirtiesContext to each test class that uses #RabbitListeners, to tell the test framework to shutdown the context.

How to execute some code after Cucumber report is built?

I use Cucumber for jUnit runner to run BDD tests like this:
#RunWith(Cucumber.class)
#CucumberOptions(
format = {"pretty", "json:target/cucumber.json"},
glue = {"com.company.bdd.steps"},
features = {"classpath:bdd-scenarios"},
tags = {"~#skip"}
)
public class CucumberTests {
}
I would like to have beautiful HTML reports from https://github.com/damianszczepanik/cucumber-reporting
And i made jUnit #AfterClass method:
#AfterClass
public static void buildReport() throws Exception {
List<String> srcReportJson = Collections.singletonList("target/cucumber.json");
Configuration configuration = new Configuration(new File("target"), "AEOS BDD Integration Tests");
new ReportBuilder(srcReportJson, configuration).generateReports();
}
The problem is that cucumber.json is empty when #AfterClass method executes. Hence i can't build pretty HTML report.
Is there any hook which i can use to execute some code after cucumber json report is already built?
PS: Cucumber v.1.1.8 is used and Java 1.7 so i was not able to try ExtendedCucumberRunner
Have you considered adding shutdown hook? Here is an example on how to add one. Code in run() method supposed to be executed before JVM shuts down.
You can take a look at custom formatter of cucumber:
Thank you for your suggestions but I just decided to use already existing Maven plugin and execute it's goal right after test goal.
wjpowell posted this suggestion in the cucumber-jvm issues:
"You don't need to do this in cucumber. Use the #beforeclass and #afterclass annotation from within the JUnit test used to run the cucumber tests. This has the benefit of running only for the features specified by the paths or tags options.
#RunWith(Cucumber.class)
#Cucumber.Options(format = {"html:target/cucumber-html-report", "json-pretty:target/cucumber-json-report.json"})
public class RunCukesTest {
#BeforeClass
public static void setup() {
System.out.println("Ran the before");
}
#AfterClass
public static void teardown() {
System.out.println("Ran the after");
}
}
"

Clear database at startup while using flyway

In my case, I'm using spring-boot with gradle and added flyway by simply putting compile 'org.flywaydb:flyway-core' to the build.gradle.
For a simulator run, which is in test, I would like to clear the database before each run. I've put a reset script in /src/test/resources/db/migration/V1.0__Reset.sql (with the real init sql-script at /src/main/resources/db/migration/V1.1__Init.sql), but receive a SyntaxException due to the reset script, which doesn't occur when I run it from the MySQL Workbench.
How can I reset or clear the database at startup?
-- UPDATE --
I've tried to use a Spring DataSourceInitializer, but it seems Flyway scripts are executed before the DS init, so it results in Hibernate Syntax error because the tables aren't found.
#Resource
DataSource ds;
#Bean
public DataSourceInitializer dbInit() throws FileNotFoundException, URISyntaxException {
public DataSourceInitializer dbInit() throws FileNotFoundException, URISyntaxException {
DataSourceInitializer re = new DataSourceInitializer();
re.setDataSource(ds);
re.setEnabled(true);
String str = "classpath:sql/V1.0__Reset.sql";
URL url = ResourceUtils.getURL(str);
org.springframework.core.io.Resource resi = new PathResource(url.toURI());
// new org.springframework.core.io.ClassPathResource(str)
re.setDatabasePopulator(new ResourceDatabasePopulator(resi));
return re;
}
Go for Flyway.clean(). It does exactly what you want. No need to write your own reset script.
You can use ApplicationRunner to run just after the startup and inside it do whatever you want with flyway. You'll also probably want to run migrate after clean:
#Component
public class CleanDatabase implements ApplicationRunner {
#Autowired
private Flyway flyway;
#Override
public void run(ApplicationArguments args) throws Exception {
flyway.clean();
flyway.migrate();
}
}