Antoine_Sauteron_1266

Nested Workflow Block - Abort Signal Problem

Discussion created by Antoine_Sauteron_1266 on Nov 23, 2017
Latest reply on Nov 27, 2017 by Peter WIRFS
This question came up in a recent support inquiry :
We need to use nested workflows which include 3--4 inner workflows. For monitoring, we need to use block+abort signal property in task properties so we can see from top if a job blocked the workflows.
But the problem is when we solve the problem and restart the job , inner workflow finished with ended-ok but outer workflow still stays blocked *status.
So is there any way to see the blocked jobs from upper levels of workflows or can wee continue upper workflows after restarting the jobs in them?

I don't think it's possible out of the box, therefore I'm posting the question here in hope of getting inputs from other users.


Simple scenario that demonstrates the issue:

JOBS.UNIX.BORN.TO.FAILis designed to end abnormally:

lkyd4xwwq0dn.png


It's nested in JOBP.WITH.FAILING.UNIX.JOB. The 'End' block of the workflow has a dependency stating that the job should end OK, else block and send abort signal to parent.

reexpw7ojbft.png

JOBP.WITH.FAILING.UNIX.JOB is nested in JOBP.PARENT.


Behavior :

When the Unix job fails, JOBP.WITH.FAILING.UNIX.JOB is also blocked.

If a "block" dependency is set : unblocking the job will also unblock the workflow.
If "block+abort" dependency is used : every task that was blocked in the process will also have to be unblocked.


Expectation :

Is there a way to let the top workflow know that one of its nested tasks has failed, and have the processing take over after unblocking only the affected task/job ?

Best regards,
Antoine

Outcomes