Scheduled job executes on both nodes

• Type: Bug
• Status: Closed
• Resolution: Fixed
• Affects Version/s: 6.2.10 EE GA1, 6.2.X EE, 7.0.0 Alpha 1
• Fix Version/s: 6.2.X EE
• Component/s: Fault Tolerance, Fault Tolerance > Clustering Framework, Portal Services > Scheduler
• Labels:
  • QA-CT
  • QA-NA
  • ee-ts
  • liferay-fixpack-portal-86-6210
• Branch Version/s:
  6.2.x
• Backported to Branch:
  Committed
• Fix Priority: 5
• Git Pull Request: https://github.com/brianchandotcom/liferay-portal-ee/pull/12156

Steps to reproduce
1) enable clustering cluster.link.enabled=true
2) unzip sample-scheduler-portlet into liferay sdk
2) open up sample scheduler liferay-portlet.xml and set the time to be 3 min in the future (enough time to get through the deployment
3) build and deploy the sample scheduler portlet (you may need to comment out license=ee requirement in plugin-package-properties)
4) confirm that at the designated time, the job executes by printing out a message to the log
5) use script to change between slave and master

for master you have to set up a cluster with two nodes and have the nodes change between master and slave by taking them down and up again
if you are using an IDE and a breakpoint to call initMemoryClusterJobs() before the scheduled jobs are deleted in ClusterSchedulerEngine then you can use the following script

import com.liferay.portal.kernel.cluster.ClusterMasterExecutorUtil

executor = ClusterMasterExecutorUtil.getClusterMasterExecutor()

executor.notifyMasterTokenTransitionListeners(false)
executor.notifyMasterTokenTransitionListeners(true)

script for 6.2

import java.lang.reflect.Proxy
import com.liferay.portal.kernel.scheduler.SchedulerEngineHelperUtil
import com.liferay.portal.messaging.proxy.MessagingProxyInvocationHandler
import com.liferay.portal.scheduler.ClusterSchedulerEngine
import com.liferay.portal.spring.aop.ServiceBeanAopProxy

schedulerEngine = SchedulerEngineHelperUtil.getSchedulerEngineHelper()._schedulerEngine

while (schedulerEngine.getClass() != ClusterSchedulerEngine.class) {
	invocationHandler = Proxy.getInvocationHandler(schedulerEngine)

	if (invocationHandler instanceof ServiceBeanAopProxy) {
		schedulerEngine = invocationHandler._advisedSupport.getTargetSource().getTarget()
	}
	else {
		println "Failed to unwrap class: " + invocationHandler.getClass()
		break
	}
}

if (schedulerEngine.getClass() == ClusterSchedulerEngine.class) {
	masterAddressString = schedulerEngine.getMasterAddressString(false)

	schedulerEngine.masterToSlave(masterAddressString, false)

	println "Switched from master to slave"

	schedulerEngine.slaveToMaster()

	println "Switched from slave to master"
}

---

Expected behavior: since the job has already been run, there should not be any additional messages in the log saying that a job has been executed
Actual behavior: a new message prints in the log indicating that a new job has been executed