Basic SAP NetWeaver Installation Steps (NW 640)

1. Before that download the installation guide from the http://service.sap.com/instguides and read it properly before the installation is started.
2. Prepare the OS as described in the installation Guide.
3. Install ABAP Central Instance (CI)
4. Install Database (chose the number of parallel process at an optimal level (which can be maximum up to 8, default is 3). It depends on the system resources (processors and memory in the Server).
5. Install additional Dialog Instances ABAP
6. Client Creation (using Tx Code SCC4) (UID/Password : sap* / pass)
7. Client Copy from 000 to a new client. (There will only be 2 clients initially after a fresh installation. 066 and 000).
8. Install an Add-on Java to Central Instance (CI) to the newly created client (java has no client concept in JAVA Stack), here it will install the Central Service (CS) instance (MS, ES) and also will install Dispatcher, Server process and SDM.
9. Install Add-on Java to other ABAP Dialog Instance (It will install only Dispatcher and Server process) also note that only one Server process (server0) will created when the instance is initially installed, for scaling up a new server process (server1) can be added separately after installation.
10. Decide on the Option of User Management, it could be either centrally or distributed.
Central – can be handled in the ABAP stack. (using Tx Code SU01 / PFCG) and distributed would use Java’s User Management Stack (UME) for user management.
11. Changes to JVM parameters to the Java Stack (dispatcher, server process(es)) are made after installation. The first step in the post installation steps is to run the template Configurator. It will set the SAP recommended JVM parameters setting for Java Stack.

Template Configurator is found in the path :
(/usr/sap/SID/SYS?global/TemplateConfig/cfgtemplategui.bat) or http://hostname:javaport/nwa ->Deploy and Change.
It is done individually for each of the instances immediately after the ABAP installation.


Is it possible to add a Java Stack alone, without ABAP in a ABAP+JAVA installation scenario. ?
In an ABAP+JAVA stack combined (java add-on to ABAP stack) installation, it is not possible to install java stack (individually, without ABAP stack) in any instance or in any host. So any additional instance of Java Stack can only be installed as an add-on to a fresh ABAP stack .i.e. both have to be installed together. For scaling up java stack a further server process can be added, BUT ABAP stack alone, without JAVA Add-on, can be installed.
A new Server process can be added by Logging into “ConfigTool” "Current Instance" -> Select the Option “Server” -> “Add”


Difference in 6.40 and 7.00 installation steps.
In 7.0 we have to choose a particular installable component during initial steps based on the Client requirement (installation scenario). ie. If the client’s requirement is of installing EP, the nEP alone will be selected for installation. Since EP requires only NW Java Stack, it will install only Java Stack with EP component. Further it will combine CI and DB installation in one step, while 6.40 had two separate steps for CI and DB installation. In ABAP+JAVA Stack installation, In NW 7.0 it will automatically create three clients 000, 066 and 001 and java stack will be automatically added to 001 client. While in 6.40 ABAP will be installed first then a fresh client has to be created and then JAVA stack is installed as an add-on to the fresh client.

Kindly note that In 7.0 after installation if the Client to which Java Stack is added as an add-on, has to be changed, then it is possible by making a few changes. The changes can be made by using Visual Administrator to change the Client Number.

How to Activating emergency user (Super Administrator User (sap*)) in JAVA Stack

Go to Config Tool -> click on the button “Switch to configuration edit mode”
Expand the tree :-
“Cluster_data -> Server -> cfg -> Services -> Propertysheet com.sap.security.core.ume.service”

When the “Super Administrator” parameter is activated to true and the password being reset, all other users will locked.
Set the parameters
ume.superadmin.activated = true
ume.superadmin.password = a new password


Restart the Java Cluster...

Logon as sap* and new password, correct the problem. then revert the change made to the parameterume.superadmin.activated to false and Restart the Java Cluster.

How to add a new Server Process to a Java Instance

A server process is added to scale up your Java Instance. In an ABAP+JAVA stack combined (java add-on to ABAP stack) installation, it is not possible to install java stack (individually, without ABAP stack) in any instance or in any host. So any additional instance of Java Stack can only be installed as an add-on to a fresh ABAP stack .i.e. both have to be installed together. For scaling up java stack a further server process can be added, BUT ABAP stack alone, without JAVA Add-on, can be installed.

Start the SAP‘s “Config Tool”
Identify the instance where you have to add Current Instance -> Select the Option “Server” -> “Add”

SAP Java requires 4 Systems Landscape Architecture while ABAP requires 3 Systems Landscape Architecture, Why ?

ABAP Stack require 3 System Landscape
ABAP requires 3 systems Landscape because ABAP modifies programs and moves them as a program to the consolidation for Testing. Since ABAP has both have Client Dependent and Client Independent data(s) and programs (cross client data consists of Programs and Data Dictionary), modification to such programs cannot be tested reliably in the same (development) system, which anyway would work correctly. So to simulate a production environment the changes need to be tested in Quality System. So an ABAP system would require 3 and not 2 systems landscape.

JAVA Stack require 4 System Landscape
In a Java stack the program are not moved alone, they are moved along with its other dependencies as a Software Component (jar, ear files). In the development system a user can copy the program and modify it and but would require a consolidation system for consolidating and building the application (Software Component). So a program is transported from Development System to Consolidation System. The Consolidation system will build the application. Once the software components are built they would be moved to Test System for Testing and finally to delivery System (Production System). Thus a Java Stack would require a 4 system landscape for a development Infrastructure.

In addition to these 4 Servers, there would be one more system which will host the Netweaver Development Infrastructure Application (NW + NWDI)

Development System -> Consolidation System -> Test System -> Productions System


Development : will have the modified Sources
Consolidation : Will consolidate the modified program from Development system and will be built as a Software component (jar File)
Testing : Software Component will tested here
Production : Productions System

Product -> Software Component -> Development Component -> Program


SAP NW 2004s Development Infrastructure
http://help.sap.com/saphelp_nw04s/helpdata/en/71/0369404c65587ee10000000a155106/frameset.htm

How To enable STATTRACE in Java Stack

Difference between STATTRACE and DSR
DSR will give time spent for the complete job, while STATTRACE will in a very detailed with detailed time spent in each of activity. So generally first a basic function is monitored using DSR. When something is noticed which has to be drilled down in detail, then STATTRACE is used to monitor in detail.

To enable STATTRACE in Java Stack, go to Visual Administrator and then go to the tab
Runtime -> Trace Config. Here activate the traces


Then go to the tab Application Tracing and restart the node to be traced in the “Maintenance Mode”.


The results can be viewed in the ABAP Stack in the tx Code ST03G


And the detailed result can be viewed in the Tx Code STATTRACE.