SAPTechArena

Unicode : Multi language support

Non Unicode : single language support 

Unicode supports all multiple languages like english, chainaa, german,..etc.
Nonunicode :only supports  two languages i.e english and german.

After NetWeaver and ECC 5.0 and etc. supports unicode kernel.
Unicode provides a unique number for every character, no matter what the platform, no matter what the program, no matter what the language.
 
OLD Code pages have the following disadvantages!

• They cover only a subset of all characters. 
• Different codepages have incompatibilities between each other. 
• Data  exchange is restricted between code pages. there are simply to many code pages.
•  Unicode has the cabaility to support all the languages in the world in one code page! 
• it support 65,000 characters and has room to support an additional 1 million characters.  

Note *

* Unicode is mandatory for SAP systems deploying JAVA applications.
* Many companies are adopting service oriented architecture standards providing
   web services to enable global interoperability. these standards  require Unicode.

What is saproot.sh?

saproot.sh is used to configure the ownership and permissions of some kernel files (BRTOOLS), including the SAPOSCOL. We have to execute the saproot.sh script after we have performed the kernel upgrade. The script need to be execute under root user and can be find under below location.

/usr/sapmnt/SID/exe

How to use

To execute the saproot.sh, just put the SAP system name (SID) and enter.

./saprootsh SID

Starting the SAP System

Stopping the SAP System

Pre Export Process In Application level :- 

• Check the Java JRE version (Min 1.4.1).
• Check the QCM table entries of SAP.
• Check for cancelled and update task
• Delete table entries from TATGPC, TATGPCA
• Check the incremental table conversion
• Cancel all released jobs
• Check for the operation modes
• Delete unnecessary spool data
• De-schedule all SAP jobs
• Delete all the batch jobs
• Release all repairs and correction and transports
• delete background input logs
• Delete job logs
• Delete the tRFC
• Check for the code pages installed
• Check the DDIC password for 000 client

Pre Export Process In OS level :-

• Deschedule all OS DB backups
• Shutdown all external interface
• The source should have minimum SP3

Pre Export Process In DB level :-

• Update DB statistics
• Change the DB to no archive log mode

Steps to Start the DB export

• Enter the system as adm
• Create to directory in the source system eg, Export, Install
• Check for the PSAPTEMP, the size should be 2 times the largest index 
• Start the R3setup.bat program from the kernel CD (use the 45B_EXT kernel cd for the migration as the target system is AIX 5.3, always check the PAM before using the kernel cd.)
• Edit the DBEXPORT.R3S file and skip the update of statistics as it is already done and update statistics of R3SETUP is high time consuming. It can even run for days for a DB size more than 1 TB
• Edit the DBEXPORT.R3S to stop for splitter
• Start the export
• The export stops at the given breakpoint for the splitter
• Split the STR files generated
• Repeat the export process

Import process

• Check the Java version ( Minimum JRE version 1.4.1 )
• Check all necessary drives and filesystems. ( #mkdir /oracle/client/92x_64, # cd /oracle/stage, #mkdir 920_64 )
• Create to directory eg, Install, Import
• From install directory run instool.sh of Ext. kernel cd for AIX.
• set environment.
• #csh
• #setenv LIBPATH /sapmnt//exe:/oracle//920_64/lib.
• Install the central instances. R3SETUP –f CENTRAL.R3S
• In the next installation process choose “Exit”
• Install Oracle software as per WAS640 guide with ora user.
• After the oracle installation is done repeat the installation of the R3SETUP
• The R3SETUP stops for Migration key check phase
• Create the tablespaces of the required sizes
• Restart the database in no archive log mode
• Start the import process
•  Complete the post installation steps from the migration guide of 45 B

Some of the problems faced with the above mentioned source and target system

• During the export phase after the STR files are created the export stops with an error saying 

“Cannot connect to DB”

Solution: Delete the following R3load, R3szchk, R3ldctl from the Instal directory.

•  Use update of DB statistics by sapdba or brtools. Avoid using the update of statistics by the export

Process

Solution: Edit the DBEXPORT.R3S file and delete the lines DBCOMPUTESTAT4MIG_XT*,
DBCOMPUTESTAT4MIG*, BRCONNECTEXPSTAT*

•  Export takes a huge time

Solution: Use the java based package splitter to split the STR files for the export.

•  Export fails complaining the PSAPTEMP size is small.

Solution: Check the size of the PSAPTEMP before export. It should be atleast 2 times the size of the largest index for R/3.

• During the import process check the aio settings in the OS level.

Solution: We have used the mix server of 512 and max server of 1024 during the import process

• 6. The import process fails due to lack of space to extend the extent at PSAPROLL

Solution: Try using the PSAPUNDO in place of PSAPROLL

• 7. The total system downtime could be reduced if MIGMON was used. The firewall rule prevented 

from the use of migmon between the source and the target.


Post Installation jobs for BASIS area

•  Replace the disp+work of the kernel.
•  Delete the table entries as mentioned in the Heterogeneous system copy guide of 45B
• Edit the RFC destination address of other connected system.    
•  Change the printer definition of the target. Test all the printers are working or we need to make changes in the settings.
•  Set up the TMS
• Check for the ALE and IDOCS flow after the new system is up
•  Schedule all the basic BASIS background jobs.
•  Take a full offline backup

Reason for moving to Oracle 9.2.0.7

9.2.0.4 being the first stable release of R/3 had many bugs which are listed in the SAP and Oracle sites were later removed in later subsequent releases.

This oracle will be upgraded to Oracle 10.2.0.2 and for the upgrade the source should be at oracle 9.2.0.7 before it can be upgraded to 10g. 

As we go higher up in the in the Optimizer gets improved hence the execution plans are improved and this results is transactions executing faster

Using of the Package splitter

The java based package splitter can be used to split the STR files

•  Open new terminal preparation for splitter tool.
•  Edit packet_splitter_end.properties file 
•  Use splitter guide for clarification.
•  setenv JAVA_HOME /usr/opt/java141.
•  using adm user, start split process.