Bug 790363

Summary: Dependent processes are not aborted
Product: [JBoss] JBoss Enterprise BRMS Platform 5 Reporter: Tomas Schlosser <tschloss>
Component: jBPM 5Assignee: Kris Verlaenen <kverlaen>
Status: VERIFIED --- QA Contact: Radovan Synek <rsynek>
Severity: urgent Docs Contact:
Priority: unspecified    
Version: BRMS 5.3.0.GACC: lpetrovi
Target Milestone: ---   
Target Release: BRMS 5.3.0.GA   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Tomas Schlosser 2012-02-14 10:48:12 UTC 
Description of problem:
When having a dependent process that does not have to wait for completion and the parent process is aborted, the subprocess is still in active state.

Version-Release number of selected component (if applicable):
BRMS-5.3.0-ER3

How reproducible:
Every time

Steps to Reproduce:
1. start process with subprocess (dependent, no wait for completion)
2. abort parent process
3. get subprocess state
  
Actual results:
STATE_ACTIVE

Expected results:
STATE_ABORTED

Additional info:
Pull request will follow shortly
Comment 1 Tomas Schlosser 2012-02-14 11:00:05 UTC 
Pull request #52 was submitted
Comment 2 Kris Verlaenen 2012-02-28 23:36:54 UTC 
It is correct that abortion of dependent sub-processes only works if waitForCompletion is set to true.  I'm wondering if we should just document this (as it seems reasonable to assume that you want to keep track of whether it is completed or not if you want to abort it if necessary), or whether we should attempt to implement it anyway.

My suggestion would be to document it, as it would keep the code simpler, and we already offer a workaround that is almost identical, where you could use a parallel gateway before the sub-process, if you want to trigger rest of the path (not wait for completion) while also supporting aborting as a dependent sub-process.
Comment 3 Tomas Schlosser 2012-02-29 07:30:37 UTC 
Documentation says: "Independent: If this property is true, ... ; otherwise the active sub-process will be cancelled on termination of the parent process (or cancellation of the sub-process node)."

It doesn't say anything about cancelling only processes that are waited for. IMHO the sub-process that is dependent should be cancelled even if it is not waited for. If you need an example: a long term process starts recurring sub-process (therefore it is not waited for it to complete) that should only run as long as the starting process lasts. The workaround you speak of is a possible solution, but it is just a workaround.
Comment 4 Tomas Schlosser 2012-03-21 08:00:47 UTC 
I just found out a new problem in this - the subprocess is aborted (beforeNodeCompleted event is fired saying that state of process instance is STATE_ABORTED). The real problem is that the subprocess was waiting for a signal and when the event was signalled, the subprocess (that was already supposed to be aborted) fired beforeProcessCompleted with STATE_COMPLETED.
Comment 5 Kris Verlaenen 2012-04-15 21:51:19 UTC 
This has been updated in the community docs, you should only use dependent set to "false" if "wait for completion" is set to "true".
Comment 8 Tomas Schlosser 2012-04-19 10:08:19 UTC 
This is already in BRMS BPM user guide, therefore marking the bug as verified.