Monday, September 30, 2019

Degree of Parallelization in SAP BW

Normally by SAP BW systems a large portions of data is processed. Such a data volume normally cannot be processed by single process. Depending on the available hardware resources the process that crunches the data should be parallelized. SAP BW systems similarly as any other moderns DWH systems is supporting this feature. The number of processes that are processing the data is called degree of parallelization.

In the SAP BW it is possible to define the degree of Parallelization in t-code RSBATCH for multiple different tasks/objects within the SAP BW. There are (in version 7.5 SP15) about 106 different processes. As per table RSPROCESSTYPES these includes:

ABAP            Execute ABAP Program
ADSOACT      Activate Requests in DataStore Objects (advanced)
ADSODEL      Deletion of Requests from DataStore Objects (advanced)
ADSOOLR      Delete Overlapping Requests from DataStore Object (advanced)
ADSOREM      Clean Up Old Requests in DataStore Objects (advanced)
AGGRFILL      Initial Fill of New Aggregates
AND              AND (Last)
ANPR            Execute Analysis Process
ARCHIVE       Archive Data from an InfoProvider
ATTRIBCHAN Attribute Change Run
AUTOCLASS   Automatic Classification
BIAINDEX      Initial Activation and Filling of SAP HANA/BWA Indexes
BIA_TPS        Update Explorer Properties of BW Objects
BOBJDSJOB   Start Job in SAP BusinessObjects Data Services
BOBJEVENT   Trigger Event in SAP BOBJ BI Platform for BW Data Changes
BPCADMCDM BPC: Create Dimension
BPCADMOTC  BPC: Optimize Application
BPCADMVLG  BPC: Validate Logic
BPCADTARC   BPC: Archive Audit
BPCAPPS       BPC: Application Source
BPCCLCUBE   BPC: Delete Application
BPCCLEAR     BPC: Delete BPC Tables
BPCCMCOVT  BPC: Convert for CM
BPCCOMS      BPC: Delete Comments
BPCECHAIN   BPC: Execution of Chain
BPCEPURPT   BPC: Generate Equity Pick-up Audit Report
BPCFILE        BPC: Data Connection
BPCFINOPT    BPC: Complete Optimization
BPCIFS          BPC: InfoProvider Source
BPCIOBJS      BPC: InfoObject Source
BPCLOAD      BPC: Load
BPCMADLOAD BPC: Master Data Description
BPCMDDEAL  BPC: Master Data Processing
BPCMODIFY   BPC: Modify Dynamically
BPCONLINE   BPC: Online Status
BPCOWCALC  BPC: Ownership Calculation
BPCPSAS       BPC: PSA Source
BPCTDHVER   BPC: Manage TDH Version
CHAIN           Local Process Chain
CHGLOGDEL  Deletion of Requests from Change Log of DSO (classic)
CLEARTODS   Delete Entire Content of Linked Transactional ODS
CL_ARCHIVE  Cleanup Archive
COMMAND     OS Command
COMPRESS    Compression of the InfoCube
CPS_JOB       Job in SAP CPS
DATACHANGE Trigger Event Data Change (for Broadcaster)
DBSTAT        Construct Database Statistics
DECISION     Decision Between Multiple Alternatives
DROPCUBE    Complete Deletion of Data Target Contents
DROPINDEX   Delete Index
DSOREPLIC   DataStore Object-Replication
DTP_LOAD     Data Transfer Process
EXOR            EXOR (First)
HAAP            Execute SAP HANA Analysis Process
HIERSAVE     Save Hierarchy (3.X)
INDEX           Generate Index
INTERRUPT    Interrupt Process
LOADING       Execute InfoPackage
LPI                Logically Partitioned InfoPackages
MAIL             Send Message
MDREORG     Reorganize Attributes and Texts for Master Data
MDUPDATEGK Master Data Update with Global Keys
MULTIHIER    Load All Hierarchies
NDB_MERGE  Trigger Delta Merge
ODSACTIVAT Activate Requests in DataStore Objects (classic)
ODSPROCESS Update DataStore Object Data (Further Update)
OHD_NOTIFY Open Hub Notification
OPENHUB      Data Export into External Systems (Obsolete)
OR                OR (Each)
PC_ACTIVE    Is the previous run in the chain still active?
PLSEQ           Execute Planning Sequence
PLSWITCHL   Switch Realtime InfoCube to Load Mode
PLSWITCHP   Switch Realtime InfoCube to Plan Mode
PSADELETE   Deletion of Requests from PSA
PSAPROCESS Read PSA and Update Data Target
QMSTATUS    Quality Status/Set Data Release
RDA_CLOSE   Close Request for an InfoPackage (RDA/Push)
RDA_RESET   Stop Real-Time Data Acquisition (RDA) Load Process
RDA_START   Start Load Process for Real-Time Data Acquisition (RDA)
REMOTE        Remote Process Chain
REPA_BB       Precalculation of Value Sets
REPA_BC       Precalculation of Web Templates
REPA_BP       Print in the Background
REPA_ER       Exception Reporting
REQUDEL      Delete Overlapping Requests from InfoCube
ROLLUP         Roll Up of Filled Aggregates/BWA Indexes
RS2HANA      Replicate BW Analysis Authorizations to SAP HANA
RSCNV           Remodeling Toolbox Operation
RSDASYB_PM Sybase IQ Partition Monitor
RSSDA_STAT Creation of Database Statistics for Virtual Tables
R_SALES2XI  Send POS Sales Data to XI System
SDI_INIT       Start SAP HANA EIM Remote Subscription
TB_LCCUPD   Last Customer Contact Update (Retraction)
TIMCHNGRUN Adjustment of Time-Dependent Aggregates
TRIGGER       Start Process
WODSODEL   Deletion of Requests from Write-Optimized DSO (classic)

Notice that for not all the above mentioned processes the parallelization make sense so not for all of them are the degree of parallelization is possible to set. Actually, the setting itself is done in section called “Settings for Parallel Processing” for those processes that it can be set.
The settings set in this t-code are stored in table RSBATCHPARALLEL - Control Parallel Batch Processes. Maintaining the table itself is coded in FM RSBATCH_MAINTAIN_PAR_SETTINGS.

More information:
2197654 - BW Data Extraction - Important house keeping activity in BW system

Wednesday, September 18, 2019

Find out users having a role

Just a quick tip on how to find out what are user(s) having particular role assigned in the SAP system. Normally this can be done via t-code PFCG there in role maintenance screen is a tab called Users, which shows this information.

However if one has no access to the PFCG t-code it needs it be relied on tables. The table that holds an information on assignment of roles to users is called AGR_USERS. The same info as in PFCG can be seen there.

Thursday, September 12, 2019

Dummy Source System in BW

Concept of dummy for an SAP BW system is sometimes required in case of large BW installations. The case is that there is not only connected one actual source system but there are connected several source systems that are not all completely homogeneous.

In case there are several target system for one source system it is possible to leverage option “7.0” in screen of “Conversion of source system names after the transport (view V_RSLOGSYSMAP)”. That option is only valid for BW 7.x objects like RSDS, TRFN and DTPA etc. It means that 7.x objects are imported several times (to several targets) even though they were exported only once. However in case target are not the same (not homogeneous) it may become tricky because those system may have different settings applied like different filters on DTPs, different schedules etc.

These kind of issues are addressed by introducing a source system type of dummy source system. This type of source system can only be created in development systems. It is not real source system just an alias to real one. Metadata of all depended objects are referenced this alias. During the transport remote function calls which for example are doing replication of DataSources or actually activate and run the metadata objects are re-directed to the real source system.

More information: