topic Re: Zombie threads running after stop Talend Job in trun(karaf) in Talend Studio

Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Sat, 16 Nov 2024 03:15:59 GMT

After stop Talend Job(TJ) bundle by bundle:stop command, TJ stops and writes a similar message to the log

Destroying BlueprintContainer for bundle ...
Service ... removed
Removing job ...

But threads from thread pool TJ is continued running and write to log messages.
Why? How to kill this zombies?

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 14 Feb 2020 13:42:25 GMT

When you deploy a bundle (or a feature) there are often other bundles that are created. You can see this if you use bundle:list before deploying a new bundle, then using it again after the bundle has been deployed. I suspect what is happening is that you have not stopped all linked bundles. It is hard to be sure since I'm not looking at your karaf, but that would be my first thought

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 14 Feb 2020 14:08:17 GMT

But these are threads of a my stopped TJ bundle, not system. TJ is a alone bundle with package dependencies only, not services dependency.
Ok. How I do stop this 'pool-35-thread-2'etc by karaf console?

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 14 Feb 2020 14:55:56 GMT

What does your bundle do? Is it a job, a route or a service?

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 14 Feb 2020 15:09:56 GMT

> After stop Talend Job(TJ)
Talend Job, sequentially reads from MSSQL and writes to RESTClient

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 14 Feb 2020 15:35:40 GMT

OK, can you supply your log file to show what is happening? How do you know that "pool-35-thread-2" is linked to your job?

You do realise that you do not need to run Talend Jobs in the Runtime? While you can run them there, Jobs are batch and therefore it makes more sense to run them outside of the Runtime.

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 14 Feb 2020 17:39:17 GMT

Please, see my real karaf tesb.log:

# command 'start 252'
2020-02-14T20:09:17,236 | INFO | pipe-start 252 | troller.internal.JobLauncherImpl 149 | 233 - org.talend.esb.job.controller - 6.4.1 | Adding job Sync_Bill2CRM_Test2.
2020-02-14T20:09:17,237 | INFO | pool-35-thread-2 | internal.SimpleJobTask$JobRunner 69 | 233 - org.talend.esb.job.controller - 6.4.1 | Starting job Sync_Bill2CRM_Test2
2020-02-14T20:09:47,440 | INFO | pool-35-thread-2 | m_queue_0_12.Sync_Bill2CRM_Queue 1694 | 252 - lovitmain.Sync_Bill2CRM_Test2 - 0.5.0 | Working...

# command 'stop 252'
2020-02-14T20:09:55,101 | INFO | pipe-start 252 | nt.container.BlueprintExtender$3 323 | 22 - org.apache.aries.blueprint.core - 1.8.0 | Destroying BlueprintContainer for bundle lovitmain.Sync_Bill2CRM_Test2/0.5.0
2020-02-14T20:09:55,104 | INFO | pipe-start 252 | r.internal.JobTracker$Customizer 133 | 233 - org.talend.esb.job.controller - 6.4.1 | Service lovitmain.sync_bill2crm_test2_0_5.Sync_Bill2CRM_Test2@4d2f17ca removed
2020-02-14T20:09:55,104 | INFO | pipe-start 252 | troller.internal.JobLauncherImpl 156 | 233 - org.talend.esb.job.controller - 6.4.1 | Removing job Sync_Bill2CRM_Test2.
2020-02-14T20:10:17,477 | INFO | pool-35-thread-2 | m_queue_0_12.Sync_Bill2CRM_Queue 1694 | 252 - lovitmain.Sync_Bill2CRM_Test2 - 0.5.0 | Working...

At 2020-02-14T20:10:17,477 the pool-35-thread-2 continue running, but Sync_Bill2CRM_Test2 was Resolved in karaf shell at bundle:list

Why? How I do fix it?

Thanks.

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Sun, 16 Feb 2020 19:02:30 GMT

I'm afraid I have not seen this behaviour before. However, I only run routes and services within the Karaf. What does your job do? I assume that it is a batch job and will run until it has finished, then stop. Or have you built it to run continuously? If it has been built to run continuously, this *might* be the problem. Talend jobs are meant to start and stop and not run continuously. Talend Routes are meant for realtime, continuous processing.

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Mon, 17 Feb 2020 14:49:31 GMT

My job read from DB and write to REST by http in infinite loop.
If take out a infinit loop logic to external route, a job logic is stopped as needed.
Thank you.

Re: Zombie threads running after stop Talend Job in trun(karaf)

Anonymous — Fri, 21 Feb 2020 14:27:14 GMT

Sorry I have been away for a few days. If you want to add the infinite loop back in, you can do it using a route. Keep your DB Read and REST Write in a job that just does that. Then create a Route, use a cTimer to get the route to create an exchange every second (or however often you want) and connect it to a cTalendJob that calls the job I described above. This link here should give you the idea....
https://help.talend.com/reader/pLzNMtPkhhXHQTnROXPqtQ/4mX6UhxhxfCD4Z6HV3OoCA