Different timeout for a step in workflow for different workflow instances

[Ankush_Dhingra]

Hi,

I want to set a different timeout for a participant step in a workflow for different instances of the workflow running on different payloads. For e.g. I want to set a timeout of 2 hours for 1 instance of a workflow running on payload A. And for another instance of the same workflow on a different payload, Payload B, I want to set the timeout as 3 hours.Please tell if anybody has any pointers on this.

[Sham_HC]

Utilize runmode config here.

[Ankush_Dhingra]

runmode config will not help here. It's changing the configuration parameters for a CQ instance. I am talking here about a workflow instance.

[Sham_HC]

Why not?  Unless i misunderstood your requirement.  Implement the osgi bundle with the AbsoluteTimeoutHandler & in the bundle you can get run mode parameter. 

[Former Community Member]

Hi Sham,

If I am not wrong his requirement is adding 2 participant steps (they will be in or split) in the same work flow which has 2 different timeouts. Based on the payload some times it has to take right step and some times it will go through left step. I tried this but by default it is taking latest instance timeout setting and it is not completing previous instances.

Below are the logs in time out. My understanding is, it is taking only latest timout settings per workflow.

In a nutshell can we have more than one step with different timeouts?

12.12.2013 17:57:40.376 *INFO* [pool-6-thread-2] org.apache.sling.event.impl.jobs.queues.TopicRoundRobinJobQueue.Granite Workflow Timeout Queue Stopped job queue Granite Workflow Timeout Queue

12.12.2013 17:57:40.377 *INFO* [pool-6-thread-2] org.apache.sling.event Service [QueueMBean for queue Granite Workflow Timeout Queue,1974] ServiceEvent UNREGISTERING

12.12.2013 17:57:40.378 *INFO* [pool-6-thread-2] org.apache.sling.event.impl.jobs.queues.TopicRoundRobinJobQueue.Granite Workflow Queue Stopped job queue Granite Workflow Queue

12.12.2013 17:57:40.379 *INFO* [pool-6-thread-2] org.apache.sling.event Service [QueueMBean for queue Granite Workflow Queue,1975] ServiceEvent UNREGISTERING

12.12.2013 17:58:15.128 *INFO* [pool-6-thread-4] org.apache.jackrabbit.core.persistence.bundle.AbstractBundlePersistenceManager cachename=versionBundleCache[ConcurrentCache@76a33ed9], elements=5117, usedmemorykb=4821, maxmemorykb=8192, access=94234, miss=5117

12.12.2013 17:58:15.151 *INFO* [Apache Sling Job Background Loader] org.apache.sling.event Service [QueueMBean for queue Granite Workflow Timeout Queue,1976] ServiceEvent REGISTERED

12.12.2013 17:58:15.154 *INFO* [Apache Sling Job Queue Granite Workflow Timeout Queue] org.apache.sling.event.impl.jobs.queues.TopicRoundRobinJobQueue.Granite Workflow Timeout Queue Starting job queue Granite Workflow Timeout Queue

12.12.2013 17:58:15.215 *INFO* [pool-7-thread-22] org.apache.jackrabbit.core.persistence.bundle.AbstractBundlePersistenceManager cachename=crx.defaultBundleCache[ConcurrentCache@3f73b038], elements=3252, usedmemorykb=8183, maxmemorykb=8192, access=414274, miss=115972

[Sham_HC · Accepted solution]

Hmmm...  Sorry I misunderstood run mode with workflow instance.

You can write a new class which extends from AutoAdvance and implements the AbsoluteTimeoutHandler interface[1] & in getTimeoutDate have method like [2] & can avoid OR split.

[1]   http://dev.day.com/docs/en/cq/current/javadoc/com/day/cq/workflow/job/AbsoluteTimeoutHandler.html

[2]

if (isPayLoadA()) {

  return 10 min;

} else if (isPayLoadB()) {

  return 15 min;

}

else {

  return -1; // timeout now

}