Procedure for BW on HANA migration – Export Phase
After you’ve prepared your SAP BW system properly, you can start the export activities. First SAP BW system is exported, then the export files are made available in the target system for the new BW application server.
To reduce the time of the migration process, you can start the export and import process in parallel. In this case, you must save the export files in directory which you can access in the target environment. To create the necessary directory structures, you must call the SWPM in the SAP BW original system using the EXPORT PREPARATION option.
In your SWPM system, start sapinst or sapinst.exe with a user that has administrator rights. However don’t use the <SID>adm user for this purpose. Select your system version and currently used database from the overview. Choose SYSTEM COPY-SOURCE SYSTEM – BASED ON AS ABAP TOWARDS SAP HANA DB-EXPORT PREPARATION, and click NEXT.
In the next dialog window, enter the path to the profile directory (for example, D:usrsap<SID>SYSprofile) and the password of the <SID>adm user. Then, check the SID and the server name, and click NEXT. You must now enter the export directory. Don’t use a network directory for this purpose; instead, use a local directory like D:/Export/DB.
This prevents possible inconsistencies during file export. If you want to use this option for parallel export and import, you must add a network share to this directory. Only then can you access this directory and the export files contained therein from the target system during the export run.
In the next step, choose SAP HANA Database for the target system. Then recheck all the entries before you start the execution. The SWPM creates the export directories. Most of them are empty at this point. The LABEL.ASC files and another configuration file are stored for labeling the ABAP subdirectory. The current SWPM session is completed now.
In the next step, restart the SWPM system. As done before, select your system version and the database version. Then, select SYSTEM COPY-SOURCE SYSTEM-BASED ON AS ABAP TOWARDS SAP HANA DB-DATABASE INSTANCE EXPORT.
The next dialog windows are identical to that of Export preparation phase. Enter the path to the profile directory and the password for the <SID>adm user. Then, check the system entries. Now, choose the export directory you’ve just created (e.g. D:/Export/DB). This step asks you whether the SAP BW system can be stopped later on. If you check this option, you will have more control over the subsequent processes.
In the next dialog window, enter the directory that you used for the execution of the SMIGR_CREATE_DDL ABAP report for exporting native database objects. Let’ say the path is D:/Export/SMIGR. Clicking on next gives an input screen.
Here, select SAP HANA DATABASE as the target database again. Check the option SPLIT STR files. This option allows you to specify whether the STR files can be split into various small data packages during the export process.
The STR files are structures that contain definition of tables, indices or views.
Do not select the START MIGRATION MONITOR MANUALLY option because if checked, the system does not automatically start the migration monitor (MIGMON). This monitor controls and optimizes the parallel load process.
If warnings occur, you can skip by clicking CONTINUE.
In the next dialog window, choose the NEW EXPORT FROM SCRATCH option to start a new export run. You are then prompted to select which method the R3szchk tool should use to identify the size of the original database. The database-independent option USE DATA DICTIONARY INFORMATION is selected by default here.
The second option ‘USE DATABASE-SPECIFIC IMPLEMENTATION’ may be less precise because it uses internal statistics data and is not always up to date, but is fast to execute.
After you have clicked NEXT again, an input screen opens which provides options for splitting files. If you additionally called the SPLITTING TABLE PREPARATION export phase beforehand, specify the generated file here.
You must enter general export parameters in the next input screen. Depending on the hardware used, you must specify the system-internal byte representation. SAP note 552464 provides more information on this topic. The hardware dealer of your system can inform you whether your system is little-endian or big-endian.
Also, give the number of processes that should run in parallel during the export later on. SAP recommends to give two or three parallel R3load process for each CPU.
If you want to use the option of parallel import and export process, you must specify it at this point. Select the checkbox for PERFORM PARALLEL EXPORT AND IMPORT, then click NEXT.
In the next window, you must specify the export sequence. Simply use the default setting SIZE. In this case, the export is carried out sorted by size, starting with smaller packages.
Click NEXT again, and the input screen opens. There, specify how the export data is transferred to the target system.
In the next step, you define a directory for the communication exchange in which status files are exchanged between two systems. e.g. D:ExportMIGMON. You should create this directory locally on the SAP BW application server and make it available to the target system via a network share. Check your entries once.
Next, you start the export run. Once it ends, a notification is displayed. To continue, you must shut down the SAP BW system and then click on OK button.
Don’t start the original system again
While the export is running, you can already continue with the import preparations. Ideally, however you shouldn’t start the SAP original system again. If you still need to do so, make sure that no data is loaded from the connected source systems.
If delta records are loaded while export, these will not be available in the SAP BW on HANA system and will lead to inconsistency in data. You can only resolve this using re-initialization.
Note: You can run the program UMG_R3LOAD_RUNTIME_PREDICTION to determine the approximate time needed for the export. (SAP note 857081).