The purpose of this document is to explain to the end user the process of an installation and execution of the BW Fitness Test.

Installation
Prerequisites to run BWFT successfully
DB statistics should be up to date
Query statistics should be turned ON prior to the BWFT execution
InfoProvider statistics should be turned ON prior to the BWFT execution
Steplist

Installation

Steps to install the transport requests:

1. The installation files should be copied into a transport directory. Normally, the 'cofiles' should be copied into /usr/sap/trans/cofiles and the 'data' should be copied into /usr/sap/trans/data.
2. Then the transports can be added to the Import Queue using the menu in STMS (Extras->Other Request->Add).
3. Then the transports can be imported in the order, which is described in the transports.txt file. You obtained this file in the zipped folder, which you received from a DataVard consultant.
4. User executing BW Fitness test should have this role assigned - /DVD/BWFT. It is part of the given transport.

Supported BW versions: 7.00 – 7.51 with up-to-date SPs
Supported database systems:

• MaxDB
• MSSQL
• DB2
• DB4
• DB6
• HANA
• Oracle
• Sybase ASE
• Informix

Prerequisites to run BWFT successfully

DB statistics should be up to date

• Details: In order to get accurate results for most of the BWFT collectors, database table statistics of all tables that are analysed must be up to date. DataVard BWFT analyses sizes and row counts of the database tables used for SAP system from its database statistics. Current and most accurate row count and size of a table cannot be retrieved during the BWFT execution due to performance reasons.
• Impact: If DB table statistics are not up to date, results don't have to be accurate. Outdated statistics on RS* tables can also cause performance issues during the BWFT execution.
• Action: Please check with your basis team if DB statistics can be updated short before the BWFT execution. Database statistics can be checked in DB20 transaction.

Query statistics should be turned ON prior to the BWFT execution

• Details: DataVard BWFT analyses a query and load performance in a BW system. In order to see runtime of queries executed in the past it is necessary that Query statistics (RSDDSTAT) are turned ON before the BWFT is executed. Our recommended period for the analysis of queries is 6 months, however capturing a month is sufficient. It is required that for analysed objects Query statistics must be turned ON with the setting - OLAP Statistics Detail Level = 2 - All (a system records all data from an area for the front end and calculation layer as well as data from an area for an aggregation layer and aggregation information).
• Impact: If Query statistics are turned OFF or turned ON only for selected queries this can cause some KPIs cannot be calculated properly. This applies also if query statistics are not turned on with Detail level "2 - All". This impacts only a single KPI of the BWFT - Query Performance. Query Performance KPI is not a showstopper for executing the rest of the BWFT.
• Action: Please check with your basis administrator or check the setting in a transaction RSDDSTAT.

InfoProvider statistics should be turned ON prior to the BWFT execution

• Details: DataVard BWFT analyses load performance in a BW system. In order to see details of executed InfoPackages in the past it is necessary that InfoProvider statistics (RSDDSTAT) are turned ON for all InfoProviders before the BWFT is executed (OLAP Detail level is not important here). Our recommended period for the analysis of InfoPackages is 6 months, however capturing a month is sufficient.
• Impact: If InfoProvider statistics are turned OFF or turned ON only for selected InfoProviders this can cause some KPIs cannot be calculated properly. This impacts only a single KPI of the BWFT - InfoPackage Performance analysis. This KPI is not a showstopper for executing the rest of the BWFT.
• Action: Please check with your basis administrator or check the setting in a transaction RSDDSTAT.

Steplist

1. Run the transaction /DVD/BWFT.
2. Create Run ID: Enter a Run ID (max 9 characters) and click Create.
   Recommendation: YY<abbreviation_of_company>##

Figure 1: Create Run ID

3. Double-click on Initialization

Figure 2: Main Menu on non-HANA system

Figure 3: Main Menu on HANA system

4. Fill the field Customer name with the name of your company. Do not change the values in "Time characteristics for analyses". After filling the field Customer name click on the Execute button. Then navigate back to the main screen.

Figure 4: Customer name & Time characteristics

5. Prerequisite check controls if system is ready for BWFT execution. There are three checked areas:

1. 1. Query statistics settings
   2. InfoProvider statistics settings
   3. Oracle statistics (only for Oracle DB)

Figure 5: Prerequisite check on Oracle DB

Query statistics settings:
Checks if all queries have proper Query statistics' setting – Detail level 2. If it's not the case for all queries, these are listed. You can check and change these settings in transaction RSDDSTAT. Please check Prerequisites part in the beginning of the documentation to check the impact.

Figure 6: Successful prerequisite check on non-Oracle system

InfoProvider statistics settings:
Checks if InfoProvider statistics are turned ON for all InfoProviders. If this is not the case, these are listed. You can check and change these settings in transaction RSDDSTAT. Please check Prerequisites part in the beginning of the documentation to check the impact.

Figure 7: InfoProvider statistics settings

Oracle DB:
If BWFT is executed on Oracle database, Prerequisite check contains Queries' statistics and DB statistics check as well. User can check which tables have their DB statistics out of date. If some tables have yellow/red icon assigned, the DB statistics of these tables should be updated. The up-to-date DB statistics are crucial for valid execution of BWFT. Please check Prerequisites part in the beginning of the documentation to see the impact.

Figure 8: DB statistics check on Oracle system

6. Once the prerequisite part is completed, double click on Main collectors. Set the Number of jobs. We recommend using 3 background processes. The run has no impact on the performance of the system. Leave field "App. server name" empty and "Keep alive" checked as seen below.

Figure 9: Main collector pop-up (number of background jobs)

7. You can monitor the running analyses by clicking on the monitor icon on the main screen.

8. When all the collectors from the group Main collectors are done, except for the Number range buffer collector (as can be seen in the picture below), you can continue according to point 10. If you execute HANA BWFT, continue with the point 9. Collector Number range is scheduled automatically to run again on fifth day to reach 100%.

Figure 10: Finished Main collectors on non-HANA system

 
Figure 11: Finished Main collectors on HANA system

9. If executing HANA BWFT, execute HANA post-processing task now.

Figure 12: Finished HANA post-processing task

10. Back in the main menu, please execute the step Dependent collectors. At the prompt, confirm that you would like to continue.

Figure 13: Confirm continuation

And in the settings box, enter similar values as in step 6.

Figure 14: Dependent collector pop-up (number of background jobs)

11.When all the collectors from the group Dependent collectors are done or done except for the item Biggest InfoProviders in three runs (as can be seen in the picture below) you can continue according to point. Collector Biggest InfoProviders in three runs is scheduled automatically to run on third and fifth day to collect the data about InfoProviders' growth in this period.

Figure 15: Finished Dependent collectors

12. Additional collectors' folder contains collectors which are not obligatory to execute, but are suitable for SLO project analysis. If you are not interested in organizational structure overview, you can skip this collector section and move to next steps. For detailed description of this section check Information button. If you execute BWFT on HANA system, please proceed directly to point 16 now.

Figure 16: Additional collectors

13. Back in the main menu double click on Generate tasks for DIMID collector. At the prompt, confirm that you would like to continue.

Figure 17: Confirm continuation

In the next screen click Execute. When the tasks are generated successfully, navigate back to the main menu.

14. Double click on DIMID collector tasks. Set the Number of jobs. We recommend using 3 background processes. The run has no impact on the performance of the system. Leave App. server name empty and Keep alive checked as can be seen in the picture below.

Figure 18: DIMID collector pop-up (number of background jobs)

15. When all the collectors from the DIMID collector tasks are done, continue with point 16. Sometimes the status displays a flash icon ; this means that a processed object was not accessible due to for example, data loading.

16. Make sure all of the Main Collectors and Dependent Collectors have completed their runs and scheduled jobs. This should not take more than 5 days (also with Number range and Biggest InfoProviders in three runs collector, which are executed in multiple runs). Once these jobs have run, just double click on Download analyses results into XML format in the main menu. The pop-up will ask you if you want to continue. Click Yes. Fill the File to save field with the path on your hard drive and specify the name of the file (e.g. results.xml…note that the extension .xml has to be entered as well). Leave package size as can be seen in picture below. When the XML files are generated, provide these to a DataVard consultant.