Quartz Cluster recovery mechanism - quartz-scheduler

Question

I run a simple controller with spring to test quartz capabilities.
#PostMapping(path = ["/api/v1/start/{jobKey}/{jobGroup}"])
fun start(#PathVariable jobKey: String, #PathVariable jobGroup: String): ResponseEntity<String> {
val simpleJob = JobBuilder
.newJob(SampleJob::class.java)
.requestRecovery(true)
.withIdentity(JobKey.jobKey(jobKey, jobGroup))
.build()
val sampleTrigger = TriggerBuilder
.newTrigger()
.withIdentity(jobKey, jobGroup)
.withSchedule(
SimpleScheduleBuilder
.repeatSecondlyForever(5)
.withMisfireHandlingInstructionIgnoreMisfires())
.build()
val scheduler = factory.scheduler
scheduler.jobGroupNames.contains(jobGroup)
if (scheduler.jobGroupNames.contains(jobGroup)) {
return ResponseEntity.ok("Scheduler exists.")
}
scheduler.scheduleJob(simpleJob, sampleTrigger)
scheduler.start()
return ResponseEntity.ok("Scheduler started.")
}
#PostMapping(path = ["/api/v1/stop/{jobKey}/{jobGroup}"])
fun stop(#PathVariable jobKey: String, #PathVariable jobGroup: String): String {
val scheduler = factory.scheduler
scheduler.interrupt(JobKey.jobKey(jobKey, jobGroup))
val jobGroupNames = scheduler.jobGroupNames
logger.info("Existing jobGroup names: {}", jobGroupNames)
return scheduler.deleteJob(JobKey.jobKey(jobKey, jobGroup)).toString()
}
Then I start two applications on different ports with the same code and start playing with it. Let's call them APP1 and APP2
I use PostgreSQL as JobStore.
So I run several scenarios.
1) Create the job with group1 and key1 in APP1
2) Try to create a job with group1 and key1 in APP2. - it gives the error that the job already started. The behavior is like I expected.
3) Stop APP1. I expect that the job will be executed in APP2, as it still exists in JobStore, but it didn't. Do I need to provide some additional configuration?
4) Start APP1, also nothing happens. Furthermore, the record for group1 and key1 still presented in DB and can't be started.
Do I need to modify shutdown behavior to remove job on the application shutdown and start jobs in another application? or I just need to configure the trigger in some another correct way?

Answer 1

My bad, that was a silly problem. I forget to start sheduler in my application
#Bean
open fun schedulerFactory(): SchedulerFactory {
val factory = StdSchedulerFactory()
factory.initialize(ClassPathResource("quartz.properties").inputStream)
factory.scheduler.start() // this line was missed
return factory
}