Performance problem on SWWCNTP0 tables

Fri Jul 7 15:36:05 EDT 2006

Hi Ritapac,
Mike could not have said it any betters.  I agree with him, by far the 
most likely source of performance degradation will be seen from 
outstanding workitems.  I am curious how many workitems by age are 
outstanding on that system?
Also you will probably find that an analysis by workflow process might 
yield some clues,  i.e what is the breakdown of outstanding work items by 
workflow? You could also then look at the frequency of any workflows that 
jump out of your analysis.

By the way in our environment we are archiving and have satisfied the 
requirement of record retention since you are still able to retrieve the 
wf log via GOS.  Are you saying that 15million entries represents 6 months 
worth of workflow activity?.

good luck,

Edwin.

Mike Pokraka <asap at workflowconnections.com> 
Sent by: sap-wug-bounces at mit.edu
07/07/2006 11:24 AM
Please respond to
"SAP Workflow Users' Group" <sap-wug at mit.edu>

To
"SAP Workflow Users' Group" <sap-wug at mit.edu>
cc

Subject
Re: Performance problem on SWWCNTP0 tables

Hi,
RSWIWADO has been removed from the menu path a couple of versions ago 
since the actual number is meaningless without knowing your system size 
and a number of other factors. However 15 mill is a lot but not that 
huge that it should cause major headaches unless you have a small SAP 
system.

I've also found that more important than the absolute number of entries 
is the number of open items. The status is part of most indexes and 
selections. So if your flows do not complete, that's going to cause 
performance bottlenecks. Do some snapshot analyses using SE16 of old 
items - e.g. how many WI's (select from SWWWIHEAD) in January this year, 
and how many of these are NOT in status COMPLETED/CANCELLED. This number 
should be minimal - in an ideal world everything that old should be 
complete unless you really have flows that run that long. If not, 
consider cancelling old flows.

Basis should be able to check out any bottlenecks from indexes.

Also, consider archiving workitems - tx SARA, object WORKITEM.

Cheers,
Mike

Rita Paciucci wrote:
> Hi Edwin, 
> SWWCNTP0 table has many entries (about 15.000.000) and the report 
RSWIWADO shows red symbol on it.
> We have deleted all obsolete workitem, but we can't delete every item 
because we need to save at least six months before.
> I've searched on OSS and I've found 2 notes (#702356 and #706478) but 
they don't solve my problem.
> Any idea to help me?
> Thanks a lot.
> Ritapac.
> 
> 
> ---------- Initial Header -----------
> 
>>From      : sap-wug-bounces at mit.edu
> To          : "SAP Workflow Users' Group" sap-wug at mit.edu
> Cc          : 
> Date      : Thu, 6 Jul 2006 11:32:32 -0500
> Subject : Re: Performance problem on SWWCNTP0 tables
> 
> 
> 
> 
> 
> 
> 
>> Hi Ritapac,
>>
>> How table entries are we talking about here?  Have you talked to your 
>> Basis/Database folks? Have you checked on OSS? There is a general 
>> consolidated OSS note (#706478)  on large tables that mentions table (
>> SWWCNTP0). 
>>
>> If you run the report RSWIWADO how many workflow runtime tables are 
>> showing up in green, yellow or red?  (By the way you might choose to 
run 
>> this report during an off-peak time just in case it takes a long time!)
>>
>> I hope this helps.
>>
>>
>> 
--------------------------------------------------------------------------
>> Edwin Mukusha
>> Administrative Systems Group
>> University of Nebraska
>> Phone: (402) 472-7756     Fax: (402) 472-7390 
>> Email: emukusha at nebraska.edu 
>> 
--------------------------------------------------------------------------
>>
>>
>>
>>
>> "Ritapac" <ritapac2004 at libero.it> 
>> Sent by: sap-wug-bounces at mit.edu
>> 07/06/2006 09:13 AM
>> Please respond to
>> "SAP Workflow Users' Group" <sap-wug at mit.edu>
>>
>>
>> To
>> "sap-wug" <sap-wug at mit.edu>
>> cc
>>
>> Subject
>> Performance problem on SWWCNTP0 tables
>>
>>
>>
>>
>>
>>
>> Hello everybody,
>>    my system is 4.7 (620 - SP lev. 0050).
>> Each activity on Workitems (log, execution,..) takes long long time, so 

>> the system performance are devastated; the performance trace (ST05) get 
a 
>> very long time on select in SWWCNTP0 table, but we have done all the 
>> activity to delete completed WI, reorg WF tables, upgrade kernel, and 
so 
>> on.
>> It seems that the select on this table is not optimized, but we haven't 

>> found any solution for this problem.
>> Do you have any idea about that?
>>
>> Thanks in advance for your help.
>> Ritapac
>>
>>
>>
>>
>>
>> _______________________________________________
>> 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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/sap-wug/attachments/20060707/14642737/attachment.htm