#90 How to clean up dead process from JOC console or JobScheduler database

[David Chu]

Request: How to clean up process from JOC console or JobScheduler database
Issue: JobScheduler build-in job: file_order_sink is automatically triggered, created a lots of Linux process and exhausted all of total number of process specified in scheduler.xml. JobScheduler is stuck
Details: a job chain, that's triggered by a file at a monitoring directory, contains a clean up job: file_order_sink at the end of job chain. When the job chain is started manually without a trigger file, the file_order_sink doesn't have a file to move/remove. In this case, the file_order_sink job WILL automatically retry endlessly and exhausts all of process specified in scheduler.xml. JobScheduler is stuck and no way to process any job. Those endless retries create many Linux processes and shown in database, although those Linux processes were already terminated. is there any way to prevent file_order_sink from retrying automatically or clean up dead processes from JOC console or JobScheduler database?

[AdityaJ]

Hello David,

Thank you for your request.

As per your request here is the process of clean up of database can be resolved by following below article:

• Database Cleanup Scripts

When you are using file order sink then you don't need to add ad hoc order, it will start running once a file is found in file source folder. So if you don't want your job chain to start running then just don't put files in source folder.

For more details you can follow below article:

• How to configure File Order Sink - job chain
  

Also, a JIRA issue was there which is fixed in 1.13.9 that your will stop running if you add an add hoc order, you can follow below JIRA issue for more details:

• JOC-1041: When adding add hoc orders to a job chain with file_sinks no orders will be executed any more

Please let me know the above suggestion solves your problem or you need further assistance from my side.

Thanks and Best Regards
Aditya Jain

[David Chu]

Hi Aditya:
Thank you for your email. The main concern at my support request is the
file_order_sink job. In previous versions, such as 1.5, you could add
file_order_sink in a non-file-order job chain. Since we
upgrade JobScheduler from 1.5 to 1.13, if the file_order_sink is in a
non-file-order job chain or in a file-order job chain without a file order,
it could suck all processes specified in jobscheduler.xml file and make
JobScheduler engine be stuck. I just want to know there is anyway I could
force file_order_sind just run only once without retrying to suck all of
processes from JobScheduler engine?

Thanks & Regards
David
416-9510738

On Tue, Oct 26, 2021 at 6:37 AM AdityaJ ztaj0118@users.sourceforge.net
wrote:

Hello David,

Thank you for your request.

As per your request here is the process of clean up of database can be
resolved by following below article:

• Database Cleanup Scripts
  https://kb.sos-berlin.com/display/PKB/Database+Maintenance+Scripts#DatabaseMaintenanceScripts-DatabaseCleanupScripts

When you are using file order sink then you don't need to add ad hoc
order, it will start running once a file is found in file source folder. So
if you don't want your job chain to start running then just don't put files
in source folder.

For more details you can follow below article:

• How to configure File Order Sink - job chain

https://kb.sos-berlin.com/display/PKB/How+to+start+a+job+chain+on+arrival+of+a+file+in+a+directory#Howtostartajobchainonarrivalofafileinadirectory-HowtoconfigureFileOrderSink-jobchain

Also, a JIRA issue was there which is fixed in 1.13.9 that your will stop
running if you add an add hoc order, you can follow below JIRA issue for
more details:

• JOC-1041: When adding add hoc orders to a job chain with file_sinks
  no orders will be executed any more
  https://change.sos-berlin.com/browse/JOC-1041

Please let me know the above suggestion solves your problem or you need
further assistance from my side.

Thanks and Best Regards
Aditya Jain

---

• [support-requests:#90]
  https://sourceforge.net/p/jobscheduler/support-requests/90/ How to clean
  up dead process from JOC console or JobScheduler database*

Status: open
Group: 1.13
Labels: waiting_for_process
Created: Thu Oct 14, 2021 03:53 PM UTC by David Chu
Last Updated: Tue Oct 26, 2021 05:59 AM UTC
Owner: nobody

Request: How to clean up process from JOC console or JobScheduler
database
Issue: JobScheduler build-in job: file_order_sink is automatically
triggered, created a lots of Linux process and exhausted all of total
number of process specified in scheduler.xml. JobScheduler is stuck
Details: a job chain, that's triggered by a file at a monitoring
directory, contains a clean up job: file_order_sink at the end of job
chain. When the job chain is started manually without a trigger file, the
file_order_sink doesn't have a file to move/remove. In this case, the
file_order_sink job WILL automatically retry endlessly and exhausts all of
process specified in scheduler.xml. JobScheduler is stuck and no way to
process any job. Those endless retries create many Linux processes and
shown in database, although those Linux processes were already terminated.
is there any way to prevent file_order_sink from retrying automatically or
clean up dead processes from JOC console or JobScheduler database?

---

Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/jobscheduler/support-requests/90/

To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/