Release Strategy issues in 4.6c

Sample, Rick Rick.Sample at graybar.com
Wed Jul 2 12:21:38 EDT 2008


All,
Got some good ideas.
 
I will post the details when I get the best solution working. I already have an extended object, 
and changing the BOR attributes 'sounds' like the cleanest first attempt. i.e. no other code changes in all my methods.
But, I will have to see how it impacts my other Wfs hanging on ZBUS2012 first. 
*Sigh*, Only doing reads, but a long qualification process ahead with multiple apps using 2012. 
 
Also, the other performance issue not related to Release Strategy, I plan on opening an OSS Customer Message. 
Since we brought up another client for the upgrade, it is REALLY impacting code sync'n up 
for the RUNTIME. Gotta be some performance tuning BASIS can do to help us out! 
 
Again, thanks much and I really appreciate it!
Rick
 
 
________________________________

From: sap-wug-bounces at mit.edu [mailto:sap-wug-bounces at mit.edu] On Behalf Of Mike Gambier
Sent: Wednesday, July 02, 2008 10:56 AM
To: SAP Workflow Users' Group
Subject: RE: Release Strategy issues in 4.6c


Rick,
 
We use [swc_refresh_object self] inside the Method but it's the same thing as calling FM SWO_OBJECT_REFRESH.
 
Pity about the async. You might have to resort to a WAIT UP TO '0.5' SECONDS or something.
 
MGT


________________________________


Subject: RE: Release Strategy issues in 4.6c
Date: Wed, 2 Jul 2008 10:37:59 -0500
From: Rick.Sample at graybar.com
To: sap-wug at mit.edu



Hey Mike,
Yea, COMMIT WORK AND WAIT is not an option in one of my tasks. It's async! 
 
When you say refresh the OBJECT, you are doing exactly what? SWO_OBJECT_REFRESH? This I have not tried yet.
I'll give these and the other suggestions a tonk and determine what works best. 
 
Thanks much,
Rick
 

________________________________

From: sap-wug-bounces at mit.edu [mailto:sap-wug-bounces at mit.edu] On Behalf Of Mike Gambier
Sent: Wednesday, July 02, 2008 10:22 AM
To: SAP Workflow Users' Group
Subject: RE: Release Strategy issues in 4.6c


Rick,
 
We've had problems with this kind of thing before as well. Sometimes it seems that Workflow is just too quick for the database changes to actually take hold with the implicit COMMIT WORK statement issued inside the standard code.
 
In the end we resorted to explicit COMMIT WORK AND WAIT statements inside our Methods where we wanted to be absolutely sure the changes have happened before we proceed any further. This seems to have done the trick although it's a bit clunky and really not that elegant at all.  
 
Occasionally we also ask the BOR Object to refresh itself at the end of these Methods to keep the OBJECT-_* structures up to date when we think something important may have changed.
 
Using a Virtual attribute rather than a Database attribute would enforce a certain 'delay' in retrieving the database value again because you have to execute a discrete SELECT. This might be enough time for the V2 updates to happen in the back-end in the meantime. But then again it might not be enough of a delay if the system is under stress and lots of updates are happening at the same time.
 
I guess it all depends on how soon after a database change are you trying to fetch the new value?
 
Regards,
 
Mike GT

> Subject: RE: Release Strategy issues in 4.6c
> Date: Wed, 2 Jul 2008 09:59:57 -0500
> From: Rick.Sample at graybar.com
> To: sap-wug at mit.edu
> 
> Hi Florin,
> 
> Sounds plausible and I planned to extend the object after I figured out
> the buffering issue(s). 
> To recap:
> 1. The standard BOR bus2012 swc_get_property self 'RELEASESTATUS'
> l_status gets buffered data. Not the current data I need when I need it.
> 2. I run the SELECT... right after the swc_get_property as a test,
> without a few seconds wait and still not the current data I need.
> 3. If I do #2 with WAIT UP TO 3 SECONDS, I do get the most recent data.
> 
> Now, what you are saying is if I extend the attribute, make virtual, and
> use SELECT ..., I can bypass the WAIT?
> What is the difference between getting from BOR DATABASE attribute vs. a
> SELECT in the method? The BOR bypassing the buffers?
> 
> Rick
> 
> 
> 
> -----Original Message-----
> From: sap-wug-bounces at mit.edu [mailto:sap-wug-bounces at mit.edu] On Behalf
> Of Florin Wach
> Sent: Wednesday, July 02, 2008 8:46 AM
> To: SAP Workflow Users' Group
> Subject: Re: Release Strategy issues in 4.6c
> 
> Hi,
> 
> oh, I see. There's no problem with that on purchase requisitions
> BUS2105, as that table attribute is reading by itself using a SELECT ...
> FROM EBAN, but the same form within BUS2012 uses the function module for
> a buffered read.
> 
> To work around that, create a subtype w/delegation.
> Redefine the attribute ReleaseStatus
> Change the Source from "Database field" to "Virutal"
> Create a coding with SELECT SINGLE FRGST FROM EKKO INTO
> object-releaseStatus WHERE ebeln = object-key.
> 
> Best wishes,
> Florin
> 
> -------- Original-Nachricht --------
> > Datum: Wed, 2 Jul 2008 08:37:34 -0500
> > Von: "Sample, Rick" <Rick.Sample at graybar.com>
> > An: "SAP Workflow Users\' Group" <sap-wug at mit.edu>
> > Betreff: Release Strategy issues in 4.6c
> 
> > Hi WF's,
> > 
> > I have an issue with performance. I think! 
> > I have a Purchase Release Strategy WF I am working on. 
> > User approves his step, Wf catches the Release event and 
> > evaluates for next approver. Standard approval stuff. 
> > 
> > The issue:
> > When user approves his task, wf event triggers, then I immediately go
> > and evaluate the Rel Strategy.
> > The standard BOR swc_get_property self 'RELEASESTATUS' l_status is
> > obviously getting the BUFFERED data. 
> > If I put in a WAIT UP TO 3 SECONDS then to a SELECT with BYPASSING
> > BUFFER I get the correct value 
> > from the SELECT, and I can verify that the BOR get property is the
> > buffered data. 
> > 
> > In all my time doing WF, I have never had to work around such a timing
> > issues. 
> > This solution 'appears' to work and is in a background task, so I
> don't
> > think it will be an problem. 
> > 
> > Also, I tried using the FMs in ME01 to refresh EKKO buffers, read the
> > EKKO buffered data, etc. with no success.
> > Anyone have a more elegant solution or comments with my above
> solution?
> > 
> > Thanks much,
> > 
> > 
> > 
> > _______________________________________________
> > 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
> 
> _______________________________________________
> SAP-WUG mailing list
> SAP-WUG at mit.edu
> http://mailman.mit.edu/mailman/listinfo/sap-wug



________________________________

Play Live Search Charades and win £5000 Think you know your TV, music and film? <http://www.searchcharades.com/>  

________________________________

Play now! Get fish-slapping on Messenger <http://www.fishticuffs.co.uk>  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/sap-wug/attachments/20080702/997b833b/attachment.htm


More information about the SAP-WUG mailing list