Parallel Subflow inside a Parallel Subflow??

[Mike Pokraka]

Hi Mark, 

Florin might be right about the global parallel index, in which case it should be an OSS. 

Alternatively could there be a transactional issue? Some portion of the workflow internals that hasn’t committed yet. Try changing the lock behaviour (header > version dependent > locking properties) to LB or one of the others as this is also related.

I would also try to reproduce this with two simple WFs with the multi line elements hardcoded, this will also aid when raising the OSS.

Let us know how you get on, I’m also curious...

Cheers, 
Mike

On 4 Apr 2018, at 4:17 pm, Florin Wach (SI) <florin.wach at systems-integration.net> wrote:

Hallo Marc,

parallel subflows within a parallel one ... well, yes, could be that the index ParForEach is mistakenly stored globally in the workflow processor.

With putting a dummy background step, could possibly be helpful.
What should work is to use simpler wait steps for the correlating objects. A sub-sub-flow shouldn't have a problem with that.

That dummy step "step" is normal btw.  Don't worry

Take care
   Florin

> Am 04.04.2018 um 14:39 schrieb Mark Pyc <mark.pyc at gmail.com>:
> 
> G'day again,
> 
> After sending the initial mail it occurred to me I could test my theory by directly starting the first level Subflow directly, and as guessed the subsequent parallel subflow step initiates successfully. So I'm guessing there is a limitation (potentially a bug in release Basis 731 sp 16) that means you can't have a parallel subflow within a parallel subflow. I get that there could be concerns for endless loops spawning recursive subflows but seems a bit rude to not support it at all. 
> 
> In the screen shot below you can see when I started an instance of the subflow WS90100016 that it successfully spawned an instance of WS90100019 in step 1396. No changes to bindings or data, just this time WS90100016 wasn't a subflow. 
> 
> <image.png>
> 
> Anyone know if it should ever work or am I just pushing the boundaries of the sane? 
> 
> Have fun,
> Mark
> 
> 
>> On 4 April 2018 at 19:06, Mark Pyc <mark.pyc at gmail.com> wrote:
>> G'day Wuggers,
>> 
>> Wow, been a long time, but scratching my head again so once again come to the fount of all WF knowledge.
>> 
>> I'm helping another developer who has got to a point of confusion in a parked Logistics Inv. scenario where an Invoice is Held with reference to multiple POs and multiple items within those POs. It is Saved as Held as there are no Service Entry Sheets. The solution is to send create parallel WFs out per PO to the PO creator/requestor asking them to create Service Entry Sheets (SES) if appropriate. The desire is to have a Wait for Correlated Event that an SES has been created for each item. You can't seem to create parallel Wait for Correlated Event steps, so have put the Waits inside a Subflow.
>> 
>> The binding to the Sub-Subflow  is based on &PO_ITEMS[&_WF_PARFOREACH_INDEX&]& with the Items table indicated on the Misc tab. 
>> 
>> The system doesn't throw an error, but it doesn't create the subflow either. In the log there is a phantom step with a step type of "STEP" which I can't ever remember seeing before. Note that although there is a Node number there is no Task Number and no Workitem Number
>> <image.png>
>> 
>> Is this some weird restriction that you can't have a parallel subflow within a parallel subflow??? 
>> 
>> The WF just carries on like that Step doesn't exist, which indeed it doesn't seem to. 
>> 
>> I'm not justifying the design either from a business process or technical build, but I'm at a loss as to what is happening technically when trying to implement this. 
>> 
>> Appreciate any sagely input!
>> 
>> Have fun,
>> Mark
> 
> _______________________________________________
> SAP-WUG mailing list
> SAP-WUG at mit.edu
> http://mailman.mit.edu/mailman/listinfo/sap-wug

_______________________________________________
SAP-WUG mailing list
SAP-WUG at mit.edu
http://mailman.mit.edu/mailman/listinfo/sap-wug
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/sap-wug/attachments/20180404/f7cceb5b/attachment.html