[
https://issues.apache.org/jira/browse/SLING-3699?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14043568#comment-14043568
]
Anshika Agarwal commented on SLING-3699:
----------------------------------------
As a workaround, I am unscheduling the job just before returning the JobResult
and after it is executed, using following piece of code snippet:
Collection<ScheduledJobInfo> scheduledJobs = jobManager.getScheduledJobs();
for (ScheduledJobInfo eachJob : scheduledJobs) {
Map<String, Object> eachJobProps =
eachJob.getJobProperties();
if
(eachJob.getJobTopic().equals(AssetUpdateListener.REPLICATION_EVENT_TOPIC) &&
eachJobProps.get(AssetUpdateListener.PROPERTY_FORM_PATH).equals(formPath) &&
eachJobProps.get(AssetUpdateListener.PROPERTY_REPLICATION_ATTRIBUTE).equals(replicationAtribute))
{
eachJob.unschedule();
}
}
Should I go forward with this approach? Would this approach create a problem
once this issue is fixed in sling?
> [Job Scheduling] A job scheduled using JobBuilder does not get removed if the
> JobConsumer returns JobResult.OK
> --------------------------------------------------------------------------------------------------------------
>
> Key: SLING-3699
> URL: https://issues.apache.org/jira/browse/SLING-3699
> Project: Sling
> Issue Type: Bug
> Components: API
> Reporter: Anshika Agarwal
> Priority: Critical
> Labels: scheduler, scheduling
>
> Behavior:
> If we schedule a job to be executed at a specific time for a topic using:
> JobBuilder jobBuilder = jobManager.createJob(EVENT_TOPIC);
> jobBuilder.properties(props).schedule().at(activationTime).add();
> It first creates a node of type "slingevent:TimedEvent" in crx repository
> under "/var/eventing/scheduled-jobs" and when the exact time arrives (in this
> case activationTime), a node of type "slingevent:Job" gets created under
> assigned jobs and posts an event to the JobConsumer.
> Issue:
> When the JobConsumer returns JobResult.OK the node of type
> "slingevent:TimedEvent" (scheduled job) does not get removed from the list.
> Due to this issue, when the server restarts, all the jobs that were earlier
> scheduled and had finished with their execution, get executed again to become
> nodes of type "slingevent:Job" and again post an event to the JobConsumer.
> This happens for all the jobs scheduled for a time that is before the current
> time (when the server restarts)
> Desired:
> Although, we wish to have the scheduled jobs to persist even after the server
> restarts, but only the jobs that have not yet been executed should be
> executed after restart. This implies that all the jobs that are executed
> should be removed from the list of scheduled jobs once the JobConsumer
> returns JobResult.OK
--
This message was sent by Atlassian JIRA
(v6.2#6252)
