Error when defining successor for node

[Mike Pokraka]

Hi,
This is neither unexpected nor unusual for productive systems to behave 
different: unlike dev/QA, production systems run on multiple hosts and 
use load balancing. So PRD will always be more asynchronous than others 
and tasks overtaking each other are a bigger danger, hence events being 
the recommended route....
Cheers,
Mike

Carolyn Fuller wrote:
> Paul,
>
> There isn't a BDC error that I know of. At least I know that the  
> documents are posted successfully.
>
> I don't know why there is such a difference between the COMMIT WORK  
> timing in production and our test environments but it caused another  
> problem with one of our asynchronous tasks after going live with  
> ERP2005.
>
> I will know for sure whether removing the unnecessary workflow code  
> from the web application fixes the production problem on Friday. That  
> is the earliest this code can go into production.
>
> Carolyn
> On May 7, 2007, at 4:52 PM, Paul.Bakker at osr.treasury.qld.gov.au wrote:
>
>   
>> Carolyn,
>>
>> That's a very interesting analysis and definitely worth sharing  
>> with the
>> WUG list.
>>
>> It may well explain why an unexpected outcome occurs for the workitem
>> (e.g., it may be a BDC error!).
>>
>> But it's not clear to me why it only happens in Production. Why is  
>> there a
>> 'Commit Work delay' in Prod?
>>
>> If you have systemic differences between Test and Prod, you're in  
>> for a
>> tough time..
>>
>> cheers
>> Paul
>>
>>
>>
>> |---------+---------------------------->
>> |         |           Carolyn Fuller   |
>> |         |           <fuller at MIT.EDU> |
>> |         |                            |
>> |         |           07/05/2007 21:53 |
>> |         |                            |
>> |---------+---------------------------->
>>     
>>> --------------------------------------------------------------------- 
>>> ----------------------------------------------------------|
>>>       
>>    
>> |                                                                      
>>                                                           |
>>   |       To:       Sriharsha_Medisetty  
>> <Sriharsha_Medisetty at satyam.com>                                       
>>                     |
>>   |       cc:        
>> <Paul.Bakker at osr.treasury.qld.gov.au>                                  
>>                                         |
>>   |       Subject:  Re: Error when defining successor for  
>> node                                                                   
>>   |
>>     
>>> --------------------------------------------------------------------- 
>>> ----------------------------------------------------------|
>>>       
>>
>>
>> Sri,
>>
>> I'm working on a theory that might have nothing to do with workflow
>> per se. I was not the workflow developer when this workflow was
>> implemented at MIT and I did not know anything about the business
>> process that the workflow supports. On Thursday I discovered that the
>> vast majority of agents who terminate this step do so via a web
>> application and that the web application attempts to manually handle
>> the workflow instead of allowing the event POSTED or CHANGED to
>> automatically handle the workflow. I'm thinking the error might occur
>> if the COMMIT WORK from FBV0 takes place after the manual code that
>> manually handles the workflow. We  have experienced COMMIT WORK
>> delays in our production environment that did not occur in our other
>> environments. This led to sporadic behavior differences between our
>> test environments & our production environment with other
>> asynchronous steps.
>>
>> So I've removed the code that attempts to handle the workflow
>> manually. Since this problem does not occur in any environment other
>> than production, we won't know if my theory is correct until later
>> this week.
>>
>> Carolyn
>>
>> On May 3, 2007, at 11:59 PM, Sriharsha_Medisetty wrote:
>>
>>     
>>> Hi Carolan / Paul,
>>>
>>> I am also getting same error WL411 on 4.7 system " No successor
>>> defined
>>> for Node XXXXXXX and Result - cancelled by user"
>>> In workflow I am maintaining dead line for decision step. After
>>> reaching
>>> dead line WF has to follow the "Process obsolete" path. But WF is
>>> going
>>> to error status.
>>> I want to know this problem is due to the system configuration or WF
>>> design. When we will get this type of error.
>>> What may be root cause this one. Give me some details.
>>>
>>> Regards,
>>> Sri harsha.M
>>>
>>> -----Original Message-----
>>> From: sap-wug-bounces at mit.edu [mailto:sap-wug-bounces at mit.edu] On
>>> Behalf
>>> Of Carolyn Fuller
>>> Sent: Thursday, May 03, 2007 2:34 PM
>>> To: SAP Workflow Users' Group
>>> Subject: Re: Error when defining successor for node
>>>
>>> Paul,
>>>
>>> The result is empty!
>>>
>>> In the event trace log there was a single locked document occurrence
>>> during a period when there were dozens of these workflow errors.
>>> Documents that caused workflow to end in error were associated with
>>> events that looked normal and did not have any error conditions.
>>>
>>> Carolyn
>>>
>>> On May 2, 2007, at 6:56 PM, Paul.Bakker at osr.treasury.qld.gov.au  
>>> wrote:
>>>
>>>       
>>>> Carolyn,
>>>>
>>>> I've seen this error in the past - it was also sporadic. I found
>>>> some OSS
>>>> notes mentioning this error, but none were applicable to our system.
>>>>
>>>> Incidentally, the error message text in our (620) system for WL 411
>>>> is 'No
>>>> successor defined for node &2 and result &3', which may be a better
>>>> descriptor of what is really going on.
>>>>
>>>> As you've probably worked out, the root of the problem is that the
>>>> workitem
>>>> returns a result which is (apparently!) not one of the defined
>>>> outcomes.
>>>> The workflow doesn't know which branch to take next, so it
>>>> collapses in
>>>> heap.
>>>>
>>>> Can you see in your workflow log what the outcome was for the last
>>>> workitem? (Look under the 'Result' column in the technical log).
>>>>
>>>> Can you investigate how an unexpected outcome may occur (eg due to
>>>> some
>>>> error condition, like having a locked document)?
>>>>
>>>> cheers
>>>> Paul
>>>>