인프런 커뮤니티 질문&답변

당장 생각나는 방법은 두가지정도되는데 간단히 첫번째방법부터 알려주겠다

Job job = jobRegistry.getJob(jobName);
jobOperator.start(job, jobParameters);

하면 ㄱ ㄱ가능하다(배치5를 보고 질문준것으로 이해했다. 배치6이라면 직접 JobRegistry를 빈으로 등록해야한다. 배치5에서는 자동구성되니 주입받아 스면 된다)

배치6 강의를 보고 질문줄것으로 이해했다 💀

형도 새해 복 💀💀💀

파라미터 정상 전달 확인 했고, 잡이 정상 실행된 것도 확인 했어!

반갑다 스프링형☠️  방금까지 배치6 나머지 강의 자료를 등록하느라 정신이 없었다. 10분만 휴식을 취하겠다. ☠️ 

@Configuration
class TestConfig(
    private val jobOperator: JobOperator,
    private val dataSource: DataSource,
    private val jobRepository: JobRepository,
    private val transactionManager: PlatformTransactionManager,
) {

    companion object {
        private const val CHUNK_SIZE = 200
    }

    private val logger = KotlinLogging.logger {}

    @Scheduled(cron = "*/30 * * * * *")
    fun runtestJob() {
        val jobParameters = JobParametersBuilder()
            .addString("startTime", LocalDateTime.now().toString())
            .toJobParameters()

        jobOperator.start(testJob(),jobParameters)
    }

    @Bean
    fun testJob(): Job =
        JobBuilder("testJob",jobRepository)
            .start(testStep())
            .build()


    @Bean
    fun testStep(): Step =
        StepBuilder("testStep", jobRepository)
            .chunk<String, String>(CHUNK_SIZE)
            .transactionManager(transactionManager)
            .reader(testReader())
            .writer(testWriter())
            .build()

    @Bean
    @StepScope
    fun testReader(): JdbcPagingItemReader<String> =
        JdbcPagingItemReaderBuilder<String>()
            .name("testReader")
            .dataSource(dataSource)
            .pageSize(CHUNK_SIZE)
            .selectClause("select img_name, pk")
            .fromClause("from card_img")
            .sortKeys(mapOf("pk" to Order.ASCENDING))
            .rowMapper { rs, _ -> rs.getString("img_name") }
            .build()

    @Bean
    @StepScope
    fun testWriter(): ItemWriter<String> = ItemWriter { chunk ->
        logger.info { "${chunk.items.size}개 전달받았습니다." }
    }
}

이 코드를 실행하고 나는 30초마다 x개 전달받았습니다. 라는 로그가 남길 기대한다.

하지만 로그는 아래와 같이 최초 1회만 로그가 남는다.
이유를 모르겠다,,! 하루종일 삽질중이다..

2026-01-04T19:49:30.143+09:00  INFO 98079 --- [ cached-async-1] o.s.b.c.l.s.TaskExecutorJobLauncher      : Job: [SimpleJob: [name=testJob]] launched with the following parameters: [{JobParameter{name='startTime', value=2026-01-04T19:49:30.024757, type=class java.lang.String, identifying=true}}]
2026-01-04T19:49:30.205+09:00  INFO 98079 --- [ cached-async-1] o.s.batch.core.job.SimpleStepHandler     : Executing step: [testStep]
2026-01-04T19:49:30.224+09:00  INFO 98079 --- [ cached-async-1] com.clip.batch.fcm.TestConfig            : 15개 전달받았습니다.
2026-01-04T19:49:30.231+09:00  INFO 98079 --- [ cached-async-1] o.s.batch.core.step.AbstractStep         : Step: [testStep] executed in 24ms
2026-01-04T19:49:30.259+09:00  INFO 98079 --- [ cached-async-1] o.s.b.c.l.s.TaskExecutorJobLauncher      : Job: [SimpleJob: [name=testJob]] completed with the following parameters: [{JobParameter{name='startTime', value=2026-01-04T19:49:30.024757, type=class java.lang.String, identifying=true}}] and the following status: [COMPLETED] in 100ms
2026-01-04T19:50:00.050+09:00  INFO 98079 --- [ cached-async-1] o.s.b.c.l.s.TaskExecutorJobLauncher      : Job: [SimpleJob: [name=testJob]] launched with the following parameters: [{JobParameter{name='startTime', value=2026-01-04T19:50:00.006744, type=class java.lang.String, identifying=true}}]
2026-01-04T19:50:00.065+09:00  INFO 98079 --- [ cached-async-1] o.s.batch.core.job.SimpleStepHandler     : Executing step: [testStep]
2026-01-04T19:50:00.069+09:00  INFO 98079 --- [ cached-async-1] o.s.batch.core.step.AbstractStep         : Step: [testStep] executed in 4ms
2026-01-04T19:50:00.089+09:00  INFO 98079 --- [ cached-async-1] o.s.b.c.l.s.TaskExecutorJobLauncher      : Job: [SimpleJob: [name=testJob]] completed with the following parameters: [{JobParameter{name='startTime', value=2026-01-04T19:50:00.006744, type=class java.lang.String, identifying=true}}] and the following status: [COMPLETED] in 31ms

@Configuration
class SampleScheduler(
    private val jobOperator: JobOperator,
    private val sampleSchedulerContentService: SampleSchedulerContentService,
    private val dataSource: DataSource,
    private val jobRepository: JobRepository,
    private val transactionManager: PlatformTransactionManager,
    private val jobRegistry: JobRegistry,
) {

    companion object {
        private const val CHUNK_SIZE = 200
    }

    private val logger = KotlinLogging.logger {}

    @Scheduled(cron = "0 0 19,22 * * *")
    fun runSampleJob() {

        val hour = LocalDateTime.now().hour
    
        val content = when (hour) {
            19 -> sampleSchedulerContentService.findFirstSchedulerContent()
            22 -> sampleSchedulerContentService.findSecondSchedulerContent()

            else -> throw IllegalArgumentException("Invalid hour for FCM scheduling")
        }
        val jobParameters = JobParametersBuilder()
            .addString("title", content.title)
            .addString("content", content.content)
            .addString("startTime", LocalDateTime.now().toString())
            .toJobParameters()
        val job = jobRegistry.getJob("sampleJob")
            ?: throw IllegalStateException("Job not found: sampleJob")

        jobOperator.start(job, jobParameters)
    }

    @Bean
    fun sampleJob(sampleStep: Step): Job =
        JobBuilder("sampleJob",jobRepository)
            .start(sampleStep)
            .build()


    @Bean
    fun sampleStep(
        sampleReader: JdbcPagingItemReader<String>,
        sampleWriter: ItemWriter<String>,
        jdbcTemplate: JdbcTemplate
    ): Step =
        StepBuilder("sampleStep", jobRepository)
            .chunk<String, String>(CHUNK_SIZE)
            .transactionManager(transactionManager)
            .listener(object : StepExecutionListener {
                override fun beforeStep(se: StepExecution) {
                    val cnt = jdbcTemplate.queryForObject(
                        "select count(*) from member where is_allow_notify = true and firebase_token is not null",
                        Long::class.java
                    )
                    logger.info { "PRE-CHECK count=$cnt" }
                }
                override fun afterStep(se: StepExecution): ExitStatus {
                    logger.info { "readCount=${se.readCount}, writeCount=${se.writeCount}, commitCount=${se.commitCount}" }
                    return se.exitStatus
                }
            })
            .reader(fcmReader)
            .writer(fcmWriter)
            .build()

    @Bean
    @StepScope
    fun sampleReader(): JdbcPagingItemReader<String> {

        val reader = JdbcPagingItemReaderBuilder<String>()
            .name("sampleReader")
            .dataSource(dataSource)
            .pageSize(CHUNK_SIZE)
            .selectClause("select firebase_token, pk")
            .fromClause("from member")
            .whereClause("is_allow_notify = true and firebase_token is not null")
            .sortKeys(mapOf("pk" to Order.ASCENDING))
            .rowMapper { rs, _ -> rs.getString("firebase_token") }
            .build()
        logger.info { "sampleReader bean created: " + System.identityHashCode(reader) }
        return reader
    }


    @Bean
    @StepScope
    fun sampleWriter(
        @Value("#{jobParameters['title']}") title: String?,
        @Value("#{jobParameters['content']}") content: String?,
    ): ItemWriter<String> {
        val itemWriter = ItemWriter<String> { chunk ->

            logger.info { "${chunk.items.size}개의 fcmToken을 대상으로 multicastFcm 비동기처리 하였습니다." }
        }
        logger.info { "sampleWriter bean created: " + System.identityHashCode(itemWriter) }
        return itemWriter
    }
}

형 위 코드처럼 돌리면 처음 1회는 리더에서 쿼리 조건에 해당하는 만큼 레코드를 읽어오고 writer까지 잘 돌거든? 근데 2번째 실행부터 reader에서 0개를 읽어오네? 그러니까 당연히 writer도 동작 안하고,,

reader 빈도 스탭마다 새로 잘 생성되는 걸 확인했는데 왜 실행하면 0개 읽고 끝내버릴까,,!?
리스너에서 대상 래코드가 잘 존재하는 것도 확인했어!