TADM12 SAP has introduced a new 5 - 1 - 2 release and maintenance strategy for SAP ERP .

Duration of Mainstream Maintenance for a software release is minimum 5 years . Thereafter, Extended Maintenance is available for 1 more year at an additional fee of 2% . Thereafter, Extended Maintenance is available for 2 more years at an additional fee of 4% per year . Thereafter, the software release will enter Customer-Specific Maintenance Usage types ...: . are structuring element for SAP software on a technical level. . determine the intended purpose of a system. . are realized by installing and configuring a collection of software components. . may require other usage types in the same system to operate. Unlike other standard code pages, Unicode defines a character set that includes virtually all characters used worldwide and as a result provides consistent, global character encoding. A code page describes the assignment of one character to one hexadecimal (HEX) value. One (normal: 1 byte) code page represents 256 characters. SAP strongly recommends Unicode especially in case of new installations. New releases of SAP NetWeaver and SAP applications based on SAP NetWeaver, which are released in 2007 or later, will no longer support new installations of Non-Unicode systems. With SAP ERP 6.0 MDMP (Multi-Display-Multi-processing, which permits the application server to dynamically switch code pages according to the logon language and language keys) is not supported any longer. Unicode influences the following factors: . Load of applications on hardware (CPU & RAM) ==> about +30% to +35% CPU and +50% RAM . Size of database ==> The result depends on the usage of UTF-8 (+10%) or UTF-16 (+30% to 60%). . Performance of the database (ask db vendor ...) Scenario & Process Component List ==> http://service.sap.com/scl For PI, it is a prerequisite that no other system in your system landscape has a higher release than the PI system. network (external ports):

The load between the application layer and the database layer is considerable. As a result, SAP recommends offering a bandwidth of at least 100 megabits per second for this communication. hw database disk layout: raid 5 --> data files raid 1 --> online redo logs, configuration files SAPinst is a Java based tool and thus needs a Java Runtime Environment (JRE) to be installed. For example SAP NetWeaver AS Java 7.0 needs the full J2SE SDK (Java 2 Standard Edition Software Development Kit), also called JDK (Java Development Kit). In this JDK also the JRE is contained, which also works for SAPinst. SAPinst checks environment variable SAPINST_JRE_HOME for a valid Java runtime environment. If SAPINST_JRE_HOME is not found, SAPinst also checks JAVA_HOME. As SAP NetWeaver AS Java also needs the JAVA_HOME environment variable to be set, you can simply add the JAVA_HOME variable if SAPinst and SAP NetWeaver AS Java need the same JDK version SAPinst includes a SAPinst GUI and a GUI server, which both use a Java Runtime Environment. In a standard installation SAPinst, SAPinst GUI, and the GUI server are running on the same host. If required, you can instead perform a remote installation with SAPinst, where SAPinst GUI is running on a separate host from SAPinst and the GUI server. When you start SAPinst, SAPinst GUI and the GUI server also start. SAPinst GUI connects to the GUI server with a secure SSL connection, and the GUI server

connects to SAPinst. SAPinst uses port 21200 to communicate with the GUI server. The GUI server uses port 21212 to communicate with SAPinst GUI. To get a list of all available SAPinst properties, start sapinst with the parameter -p SAPinst is controlled by XML files. The main XML files are: . dialog.xml - Contains all dialogs used in the installation . keydb.xml - SAPinst records the installation progress and user input in the keydb.xml file . messages.xml - Contains all messages used in the installation . control.xml - Contains the component definition used by SAPinst . packages.xml - For software package administration If a slow connection (e.g. network connection) is between the DVD location and the host where SAPinst runs, it is highly recommended to copy the DVDs content to the local file system before starting the installation. SAPinst can do that copy job prior to the installation just by specifying a location in the field Copy Package To. As of SAP NetWeaver 7.0 strong encryption is mandatory for the SAP NetWeaver AS Java and for all usage types that are deployed on it. You need to obtain the Java Cryptography Extension (JCE) Unlimited Strength Jurisdiction Policy Files 1.4.2. beforehand so that they can be installed by SAPinst. Download the JCE policy files for your platform at http://java.sun.com/products/jce/downloads Normally when doing a domain installation you need to be a domain administrator. Anyhow it is possible to do a domain installation with local administration rights. In this case the domain administrator has to prepare some additional things. Caution: Do not use the user <sid>adm for the installation of the SAP system. When the Maximize data throughput for network applications option is selected, the Windows file cache is reduced in size so that application server software, such as an SAP system, has more resources available. This also decreases the time needed for the installation, especially for the database import. To check the setting go to Start Settings Network Connections, right click on the active connection and choose Properties from the context menu. If your transport host is not the central instance host, you have to create the transport directory as follows: 1. On the transport host, create the directory \usr\sap\trans. 2. Grant Everyone the permission Full Control for the directory. 3. Share the usr\sap directory on the transport host as SAPMNT. This enables SAPinst to address the transport directory in the standard way as \\SAPTRANSHOST\SAPMNT\trans. Note: These permissions are only necessary during the SAPinst installation. You must remove them after you have finished the installation. After the installation, you only have to grant Full Control on this directory to the SAP_<SAPSID>_GlobalAdmin groups of all the systems that are part of your transport infrastructure. installation checklist on WIN:

for Unix, similar to WIN but kernel parameters checked/set, setting filesystems:

To use an automatism to install the frontend software SAP GUI for Windows on all end-user computers you need to use the SAP GUI Installation Server. Note that a prerequisite for using the installation server is the installation of
Microsoft .NET Framework 1.1 SP1. cscript %SystemDrive%\inetpub\AdminScripts\adsutil.vbs set w3svc/AppPools/Enable32bitAppOnWin64 1 (Execute this command after the installation of .NET framework)

SAP GUI installation server can only trigger the installation and update of SAP GUI for Windows. Java licensing In case you have a system based on SAP NetWeaver AS Java (without ABAP), for example an SAP NetWeaver Portal system, you have to install the SAP license with the SAP Visual Administrator. Log on to the Visual Administrator. You may need to create a new connection entry first. For this you need to know either the HTTP port of the Java Message Server (default 81<instance number>) or the P4 port of a Java Dispatcher (default 5<instance number>04). In the Visual Administrator go to the server node of the central instance and select the Licensing Adapter Service. Copy the Hardware Key to request the license from SAP. Afterwards choose Install license from file to install the license. post installation full backup for Windows 1. Save the Registry. 2. Save system state data. 3. Back up all SAP-specific and database-related

directories. Online Help (SAP LIbrary) For a given front-end platform, a setup variant (SPRO, Sap NW, General settings, Setting variants for Help-Sap Library = SR13) specifies a help type, the location of help files (server and/or path), and the language version of help files types: . Compiled HTML files for help type HtmlHelpFile (can be viewed directly from the DVD in Microsoft HTML Help Viewer) . Standard HTML files for help types PlainHtmlHttp and PlainHtmlFile (are stored in a packed format in the archive Plainhtm.xxx on the DVD) . DynamicHelp: standard HTML format for its files; the file access takes place using the Knowledge Warehouse server. - three steps to install SAP online documentation: Install the help files. Customize setup variants for online help. Install a web browser or viewer. By choosing SAP Library in the Help menu, you can display the online documentation. For SAP NetWeaver, for example, the SAP Library currently offers access to more than 10,000 documents. . If you only use front ends with Microsoft Windows, we recommend you install the help type HtmlHelpFile. . If you have no Web server and deploy different front end platforms to Microsoft Windows 32-Bit, you have to use the help type PlainHtmlFile. . If you want to call the SAP documentation from the SAP GUI for Windows (Win GUI), the SAP GUI for HTML (Web GUI), or a browser-based SAP application, you cannot use the help type HtmlHelpFile to do so. . WITHOUT configuration in transaction SR13, even if there is a sapdoccd.ini file on the front end, it is NOT evaluated. The default message in this case is: No documentation available. Online documentation call can be controlled on Microsoft Windows front ends with entries in the file sapdoccd.ini (WIN dir or SAPGUI program files dir). If this file exists with valid settings, these override the system settings for the relevant front end. saprouttab Standard entries in a route permission table appear as follows: P/S/D <source-host> <dest-host> <dest-serv> <password> <source-host> and <dest-host> could be SAP routers. . P(ermit) causes SAProuter to set up the connection. P(ermit) entries can contain a password. SAProuter checks whether this password corresponds to that sent by the client. . S(ecure) only allows connections with the SAP protocol; connections with other protocols, such as TCP, are not allowed. . D(eny) prevents the connection from being set up. SAProuter is an SAP program that acts as an intermediate station (proxy) in a network connection between SAP systems, or between SAP systems and external networks. SAProuter controls the access to your network (application level gateway), and, as such, is a useful enhancement to an existing firewall system (port filter). You require SAProuter if you are using any of the following: . Remote Services: e.g. EarlyWatch

. Remote Consulting . SAP Support Portal (the SAP-based service system, which provides the technical link between SAP customers and SAP: SAProuter controls connections from SAP to your SAP system.) Your production client must be a copy of the SAP reference client 000. Copy the client with transaction SCCL and profile SAP_CUST (if you don't want to use 001) . steps needed to perform a (post-installation) client copy 1. Maintain the client in table T000. 2. Copy the client (access in target client with sap*/pass). 3. Check log files. SAP delivers the software with standard client 000. You may not work in client 000. But you may use client 001 as your productive, if it exists. Client 001 is a copy of client 000. However, if no client 001 exists, SAP recommends that you begin SAP system implementation by creating a new client as a copy of client 000. SMLT Supplementation actions are client-specific. The languages are supplemented in the client to which you are logged on . steps needed to install additional languages 1. Classifying the language. 2. Scheduling of language transport. 3. Scheduling of language supplementation. Activation of extension set From the technical point of view, SAP ERP Central Component Extensions are delivered and installed as part of an Add-On for SAP ERP Central Component, which is SAP ERP Central Component Extension Set. Client-independent activation switches allow you to use the functions contained in SAP ERP Central Component Extension Set components. Generally, you cannot deactivate an activated switch Java AS 7.0 Configuration Wizard When you have installed a SAP NetWeaver system, the Configuration Wizard makes technical settings which are required for the technical processing of a system or a technical scenario, e.g. connectivity, service users, usage-type initialization. You cannot use the configuration wizard after an upgrade, an Add-In installation and the enablement of additional usage types. Caution: You can run the Configuration Wizard only once and only directly after you installed and patched the system. ==> (Call http://<server>:<port>/nwa in a browser, and logon with a Java administrator user. Navigate to Deploy & Change tab and choose Configuration Wizard) Java AS conf To prevent locking the administrator in case you do change its password and forget to update the entry secure storage, we also recommend you create a second administrator user after installing an AS Java based system. java dispatcher: . Connections Manipulation . configure the maximum number of user connections that the dispatcher can handle simultaneously and a timeout for establishing these connections (Kernel-ConnectionsManipulator)

. Setting Service Load Timeout . configure the maximum time for which the services on a cluster node have to be started. (Kernel-ServiceManager) There are two different tools available to integrate an additional usage type. Depending on your existing system (installed usage type) you use either the installation tool SAPinst or Java Support Package Manager JSPM to implement additional Usage Types. Installation tool SAPinst is used to install usage type AS-Java or any AS-Java based usage type (e.g. Biller Direct) to your existing AS ABAP based system. The installation is separated in three steps (from Additional Sw Life-cycle tasks-Java Add-in for Abap): 1. Installing Central Services Instance Java Add-In 2. Installing Database Instance Java Add-In 3. Installing Central Instance Java Add-In Since SP12 Java Support Package Manager JSPM can be used to install an Java based usage type to an existing Java based system. It is necessary to download the suitable Support Package Stacks for the new usage type and the Support Package Stack List SPS##Tab.xml. To reduce time and expenses for an system upgrade, SAP introduces Enhancement Packages. SAP Enhancement Packages for SAP ERP 6.0 provide organizations, that have implemented SAP ERP 6.0, with additional high-value applications and enable them to install new applications quickly ... Without activating the deliveries for SAP ERP enhancement package, neither the functionality delivered nor the corresponding user interface enhancements can be used. Enhancement Packages are cumulative. You cannot uninstall an Enhancement Package. To be able to use the new function of the Enhancement Packages, you must activate certain business functions (SFW5). Java patching (product, sw comp <--> SP=SC, dev comp or patches=DC) . The usual file format of an SP is the SCA format (Software Component Archive, it contains a specific number of SDAs) . The usual file format of a patch (hotfix) is the SDA format (Software Deployment Archive) JSPM ==> updating, deployment (SAP and third-party software components), business packages, complete Support Package Stack. In the JSPM Gui, you can find the activated Usage Types in the Deployed Components tab. Only the definition file (SPSTab.xml) corresponding to the Support Package Stack should be present in the file system as otherwise problems may occur. The storage location of the JSPM Inbox is defined via the parameter DIR_EPS_ROOT Check the free space in the file system for the SDM file transfer directory /usr/sap/<SID>/<central instance>/SDM/program/temp (Temporarily available disk space: 1.5 times the size of the Support Package files) | You can modify the SMD's file transfer directory .. If the Support Package Stack contains a Kernel Update, JSPM starts with the Kernel Update. JSPM updates the kernel binaries of all the instances (dialog instances and CS instance), automatically stops the central instance ... The SDM (Software Deployment Manager) is a client/server application. It is responsible for deploying software in AS Java and is therefore used in close conjunction with NWDI, JSPM, etc.

On the AS Java side there runs the SDM server that is started by the central instance of the AS Java automatically. The client provides a graphical user interface which you can start using the script RemoteGui.bat or RemoteGui.sh via the file system /usr/sap/<SID>/<central instance>/sdm/program. Within the framework of software logistics the SDM (more precisely: SDM Server) is called directly from within SAP NetWeaver Developer Studio or the CMS. The Java Support Package Manager also addresses the SDM Server directly. SDM port = 5$$18 System copy You can use the following procedures/methods for the two types of system copy: . Database-specific method: homogeneous system copies only . Database-independent method: homogeneous and heterogeneous system copies You can use the SAP installation tool SAPinst to create exports and imports of the contents of your Java database, the file system and the configuration in database-independent format. SAPinst uses the tool Jload. Jload generates a database export of all the SAP objects which are defined in the Java Dictionary and archives the configuration and file system components. Release Upgrade: SAPup / SAPJup A Support Package contains all known, relevant SAP Notes that have been created since the last Support Package. It can also contain new Customizing and Customizing that has been corrected since the last Support Package. Support Packages are thus not cumulative, but are based on their predecessors. The Maintenance Optimizer is mandatory for some Support Packages and Support Package Stacks, such as the Support Packages for SAP ECC 6.0 that have been available since April 2007. Sometimes, it is essential to perform a SPAM/SAINT update before importing Support Packages (see notes). Why? For example, because a new SPAM is required to read and interpret new header information from the Support Packages. The SPAM/SAINT update also often contains minor changes to the transactions SPAM and SAINT or new functions to make them more user-friendly. You can only import a SPAM/SAINT update from client 000. import mode (With the Downtime-minimized import mode, the ABAP programs in the Import 1 module are inactive when imported.):

Since you can stop the import of Support Packages after each module, you can run the Preparation and Import 1 modules during production operation if no modifications to Dictionary objects have taken place. You can considerably reduce system downtime by selecting Downtime-minimized. This makes the import procedure as a whole longer, however, since you still need to activate the ABAP programs. The SPDD adjustment should be performed manually in each system in the landscape and not be replaced by importing a transport request. This can lead to inconsistencies and is not recommended. As of SPAM/SAINT version 7.00/0024 you can include transport requests that contain the modification adjustments from the development system into the Support Package queue for the remaining systems. This means that the modification adjustments only need to be carried out once in the development system and are then implemented automatically in the quality assurance and production systems by the modification adjustment transports. This is a similar procedure to the release upgrade. Of course, this only makes sense when the systems in a landscape are each modified in a broadly similar way. Including transport requests is totally different to importing transport requests. If you are using a SPAM/SAINT version 7.00/0024 or above, you can flag the adjustment transport requests (select Select for Transport in the transactions SPDD and SPAU). This enables you to perform the modification

adjustments in the development system alone. The Note Assistant is called with the transaction SNOTE. It has been part of the standard system since Basis release 6.10. In earlier releases, you can import it as an additional software component, an add-on. The current version of the Note Assistant can implement various types of SAP Notes: Changes to SAP programs, the creation of new SAP programs, changes to SAP function modules and several other things besides. It cannot, however, change Dictionary objects, for example. Furthermore, the Note Assistant can only change repository objects, not Customizing, for example. Changes to SAP objects that are introduced using the SAP Note Assistant do not require an SSCR key. Regular backups required for SAP NW AS Java A complete backup of all changes to SAP NW AS Java since the installation consists of: . Backing Up the Databases . Backing up the file system: Global directory \usr\sap\<SID> To ensure that no changes are made to the file system during the backup, the SDM should be stopped . Backing up the instance configuration (export by means of Config Tool) Printing in SAP two steps: First a spool request is created. The spool request contains device-independent print data and includes administrative information (such as author, date, number of copies) and the actual print data. Only when the spool request is to be output on a particular device is an output request created. The device-independent print data from the spool request is converted to the printer language that the selected output device understands. When a printout is requested, a spool work process creates a corresponding output request. This output request is transferred from the spool work process of the SAP system to the operating system spool system. The operating system transfers the data to the stored printer. A printer to be addressed must therefore be defined and addressable both in the SAP system and in the operating system. The actual document content of a spool request is stored in TemSe (for temporary sequential objects), for which you define the storage location with the profile parameter rspo/store_location. . Value db (the default value): Spool requests are stored in database table TST03 (Advantage: backup as part of the database). . Value G: Stored at operating system level in the global directory (advantage: performance). Note: You can also specify the storage location individually for the output device in transaction SPAD (menu path Edit Data Storage). - There are many ways in which a spool work process can reach an operating system spooler. These connections are described as access methods . local printing = spool work process and operating system spool are running on the same host ==> Unix: access method L (lp or lpr commands) | WIN: access method C . remote printing = the spool work process and operating system spooler are running on different hosts ==> scenarios: . Network printers provide their own operating system spoolers and are directly connected to a computer network. Printers of this type are directly

addressed from the SAP system using their names (access method U). . Access method U is used if the remote host is a UNIX system. . SAP provides the program SAPSprint for all hosts with Microsoft Windows operating systems. SAPSprint is a Windows service capable of multi-threading. Each output request is processed in a separate, isolated thread. The output requests that SAPSprint receives from the SAP system can thus be transferred to a particular printer individually. Access method S is usually used here (SAP protocol), but access method U (UNIX Berkeley protocol) is also supported. ("remote printing" access method requires communication between the spool work process and an operating system spooler over a network. We therefore recommend that this access method is only used in local networks (LAN). A spool work process that is transferring print data does not accept any other tasks during the communication. The slower the network, the longer the spool work process is blocked and unavailable for other tasks) . frontend printing The access method F is popular on Windows front ends and can and should now be replaced with access method G. Under Windows you can, for example, address the front end default printer. The printer is given a generic name in the SAP system, and is assigned to the physical device "__DEFAULT". Since the models used as front-end printers can vary considerably, the device type SWIN (or SAPWIN) is used. When this device type is used, the printer driver of the front-end computer formats the data to be printed. The front end receives the data from the spool work process, which uses saplpd.exe (access method F) or the SAP GUI (access method G) on the front end for this communication. saplpd.exe is, if necessary, started using the dialog connection between the front end and the SAP system. Front-end printing can also be used on non-Microsoft Windows front ends such as Unix and Apple Macintosh, with certain restrictions (same printer name on all front ends and identical control, such as in PostScript format or device type). SAP users can output documents on their local printers using front-end printing. These local printers do not need to be individually defined in the SAP system. Rather the system administrator only needs to create a representative output device for each operating system platform. As of SAP Web AS 6.20 for frontend printing with access method G no longer is required an SAPlpd. A Windows printer selection window is called directly from the GUI control. (Front-end printing with control technology (access method G) also enables front-end printing from the SAP GUI for Java on non-Windows-based platforms, such as UNIX. Printing using the Windows Terminal Server is also made easier with front-end printing using control technology. Of course, front-end printing requires a connection to the front-end PC. However, this method cannot be used for background processing.) If you use SAP GUI for HTML and print on your front end, you need access method F. Using this print method, the print data should be sent to the browser and displayed. You can then print the document on your front end. In this case you need to use the device type PDF1

Names: Spool server = SAP application server with spool work processes or logical server (defined in SPAD). Host printer = Name of the printer at operating system level Host name = Only for local printing, is calculated automatically from the spool server. Destination host = Only for remote printing. Name of the host on which the operating system spooler is running. . Device types:

There are drivers for the Microsoft Windows operating systems for practically all printers on the market. You can use these printers as output devices with device type SWIN. SWIN / SAPWIN cannot be used with local access methods. For many printers, you can use a generic format (such as PostScript) for which device types exist. Regular deletion of old spool requests (RSPO1041) and monitoring the consistency of the spool database (RSPO1043) are system administration tasks. External Output Management Systems (OMS) can be addressed from the SAP system using a defined interface (BC-XOM) (access method E). Since SAP Basis 4.0B (as of a particular kernel and Support Package status), you can send print output by e-mail (SapConnect) to a user (access method M). possible statuses of a spool request are Compl., Error, Waiting, Print., -, +, Proc., <F5>, Problem, and Time. Background processes - the process sequence (priority) of background jobs (for available background work processes) is actually determined as follows: . Scheduled start time . Job class (A-C) . Target server specified or not If there are no free resources available for background processing at a certain time,

but jobs are to be executed, jobs with a higher priority are processed before jobs with an earlier start time but lower priority Every job is processed without interruption by one single background work process. - differentiation between "external command" and "external program": it is possible to perform an authorization check for external commands for the command name (defined only in the SAP system). For external programs, it is only possible to set a "permitted"/"not permitted" distinction. In both cases, the program sapxpg is called at operating system level and starts the relevant program at operating system level. This program runs using the authorizations of the SAP system operating system user. - a job can be triggered time dependent or event dependent . time-dependent job scheduler runs (on instances with batch processes) with the frequency of rdisp/btctime seconds This is an ABAP program (SAPMSSY2 - an "Auto ABAP") that runs automatically in a dialog work process. As of SAP NetWeaver 7.0, the Job Scheduler also starts after a job has finished. . An event is a signal to the background processing system that a particular status has been achieved in the SAP system. An application server is specified for the processing of events that are triggered within the SAP system in profile parameter rdisp/btcname. An event-dependent job scheduler is started on this server (unlike the time-dependent job scheduler, which runs periodically on all background servers). This scheduler checks whether a job is waiting for the event that has been received. It is therefore important that the parameter rdisp/btcname contains the name of an active background server (normally the central instance). Events can be triggered in various ways: . Manually in CCMS for test purposes (transaction SM64) . Within an ABAP program (by using the function module BP_EVENT_RAISE or the RAISE method for the CL_BATCH_EVENT class) . Outside the SAP system at operating system level using the program sapevt. . You can still change a job as long as the job still has the status Scheduled or Released. - ready status: The start condition of a released job has been fulfilled. A job scheduler has placed the job in the wait queue for a free background work process. - capture: If an active job does not seem to be running correctly (for example, it is running for an unusually long time), you can terminate it in a background work process in debugging mode, analyze it, and then either release it again or terminate it completely. To do this, in transaction SM37, choose Job Capture: active job. - To set the number of reserved background work processes for class A, you can define an operation mode in transaction RZ04 (maintenance of operation modes and instances) and maintain the work process allocation for this operation mode. When doing so, you have the option of reserving work processes. We recommend that you do not reserve more than one background work process for processing class "A" jobs for each instance. - SM61: background controller list & background group of instances ...

SAPconnect (communication possibilities):

- SCOT ==> communication nodes - SAPconnect is set up client dependent - flow: 1. A message is created, for example as a WebFlow item, in the Business Workplace (transaction SBWP) or by an application. 2. The message is assigned to a node based on the selected communication type and address area stored in the queue. 3. The send process (report RSCONN01 - should run periodically in the background) starts, reads the message from the queue, and transfers it to the SMTP plug-in or an RFC destination. Mail Transfer Agent (MTA): This is a program (such as sendmail) that is responsible for forwarding and delivering e-mails. - Restrictions In the outbound direction, the SAP system always transfers the mails to one single mail server. You can set the host address and port number of this mail server. The technical background of this is that an SMTP node can have only one target connection. In order to send e-mails to different domains, the AS ABAP therefore always requires a different Mail Transfer Agents (MTA). The Application Server ABAP (AS ABAP) can receive e-mails through SMTP and redistribute these e-mails to Business Workplace (previously: SAP Office) users. In the inbound direction, the SAP system can receive mails from any number of mail servers. Each client can be reached by a separate virtual mail server (host name, port number) that is configured in transaction SICF. The best way to address the SAP system and its clients is to use a separate subdomain for each client (such as training.sap.com). These subdomains are assigned to host names and port numbers using routing rules on the mail servers. The following restriction applies: The AS ABAP cannot be an MTA itself, that is, it cannot

forward the mails to other mail servers. To avoid performance bottlenecks due to high numbers of SMTP requests (or, optionally, HTTP/S), you can use parameters to restrict the context usage in the backend SAP system for a protocol. If the following quotas are exceeded, the requests are rejected in the ICM. . icm/HTTP(S)/context_quota . icm/SMTP/context_quota For example, with the setting icm/SMTP/context_quota = 20, you specify that only 20 percent of the available contexts can be used for SMTP (mail functions). This means that you have more capacity for HTTP(S) requests if there is a high workload. In transaction SCOT, define the domain of the SAP system client under Settings Default Domain. You need to do this for the following reasons: . The SMTP plug-in logs onto the mail server with the domain as an ID. . The message ID of the outbound e-mails is compiled using this domain. . If an SAP user sends an e-mail without an Internet e-mail address, a sender address consisting of the SAP user name and this domain is generated.
Background information about the importance of default domains: Each mail requires a unique ID that has the same format as an Internet e-mail address. The AS ABAP creates the message ID from a local number and the default domain, for example, <ADR28000000001487@003.q1a.r3.sap-ag.de>. Once the SMTP connection to the mail server has been set up, the AS ABAP identifies itself with the default domain (for example, .HELO 003.q1a.r3.sap-ag.de.). However, the mail server generally has no further use for this domain and normally looks at the IP address. For these two reasons, the default domain is always required even if the sender has entered an address in the user master record.

receiving email in an Sap sys When you send an e-mail from a mail client to a user in an SAP system, the e-mail is first forwarded to a mail server and then forwarded to the SAP system by the MTA. The e-mails are sent to the mail client in the SAP system (Business Workplace) on a client-dependent basis using SAPconnect.

- incoming mail process: SAPconnect is not configured for receiving mails in the same way as for sending mails, that is, using a node in transaction SCOT. Instead, it is configured using a service in the Internet Communication Framework (ICF) (see transaction SICF). A user is also required for processing inbound mails. Use transaction SU01 to create a system user with the profile S_A.SCON. For every client of an SAP system in which inbound mails (or status notifications for sent mails) are to be received and possibly processed, you must create an SMTP server for which an assignment to a virtual mail host and logon data are stored. In transaction SICF, there is already an SMTP server for every SAP system that is shipped by SAP. Use this for the first client that you want to be able to receive mails, and create a new SMTP server for each additional client. If you are working with multiple clients, you must create a virtual host for each client in which messages are received using the SMTP plug-in. Make the following settings in transaction SICF by double-clicking the SAPconnect menu entry (see also SAP Note 546147): . On the Host Data tab page, maintain the parameter is/SMTP/virt_host_<*>= <host>:<port>,<port>,...; This parameter defines a virtual mail host for receiving mails. ...

client concept:

The system change option defines whether or not Repository objects and client-independent customizing objects are globally modifiable. If they are globally modifiable, you can specify whether or not each software component or namespace can be modified. A software component is a set of dedicated development classes (packages). Current settings

Certain kinds of customizing changes, known as data-only customizing changes, need to be carried out in a production client without being saved as transport requests. Examples of such data include interest rates, health insurance premiums, pension schemes, tax schemes, and currency exchange rates, which may require frequent adjustment in SAP systems. Since these types of changes have a frequent change rate and they are not subject to extensive testing like other customizing changes, these changes are done directly in the productive system. To avoid having to use change requests for these changes, SAP has introduced the Current Settings function. SAP-approved Current Settings for customizing objects are kept in the field CURSETTING in the table OBJH. SAP recommends that no customer changes be made to the table. When using Current Settings in a production environment:
. Client role is set to: Production . Cross-client object changes are set to: No changes to Repository and cross-client Customizing objects . Changes and transports for client-specific objects are set to: No changes allowed.

TMS A transport group consists of one or more systems that share a common transport directory. TMS supports several transport directories within a transport domain. The common practice is configuring the development system as the transport domain controller and then later moving the assignment of the transport domain controller to the production system. The system load on the SAP system caused by work on the transport domain controller is very low. There are two types of RFC destinations created by the TMS: TMSADM@<SID>.<domain name> & TMSSUP@<SID>.<domain name>. The user TMSADM is inserted as system user into the TMSADM@<SID>.<domain name> destination but there is no RFC user inserted into the TMSSUP@<SID>.<domain name> destination. Therefore, if you try to execute a function that goes beyond the limited authorizations of the TMSADM user, for example, scheduling an import, the system uses the TMSSUP destination. As a result, the system prompts you for a user name and password on the target system that does have the necessary authorizations. All repository objects are grouped into logical units called packages (formerly: development classes). The definition of each package contains an assignment to a transport layer. The transport layer is the name of a consolidation route (Delivery routes have no names). Objects, via the assignment to the package, inherit this transport layer assignment. Customizing objects (and others) are not grouped into packages. They follow the standard transport route. This standard transport route is typically named Z<SID of the development system>.

QA approval procedure When the QA approval procedure is activated, a transport request will only be imported into the delivery system(s) if all the QA approval steps are processed in the QA system and the request has been approved. Extended transport control --> transport target groups & client transport control ==> CTC = 1 (tp parameter, client dependent transport control active). External systems are like virtual systems. However, for this type of system, an separate transport directory is also defined ... The idea behind external systems is to reduce the problem of transporting between different transport domains to the problem of transporting between different transport groups in the same transport domain. The enhanced CTS (Change and Transport System) functions ( = transports of non-ABAP objects) are available as of Support Package Stack (SPS) 12 of SAP NetWeaver 7.0. For example, the Transport Domain Controller (TDC) is located on an Solution Manager System which is an ABAP+Java System. In addition, on this system, the CTS Deploy Web Service is configured. Hint: A different option is to use an AS ABAP as TDC and to connect this system to a different AS Java (different SID) on which the CTS Deploy Web Service is configured. Within transaction STMS on the TDC you can define both ABAP and non-ABAP systems. N.B.: the CTS Deploy Web Service belongs to the software component DI_CMS, so in addition this software component needs to be deployed on the AS Java part of every CTS system. It is intended however to switch this Web Service to a software component that belongs to the AS Java itself with a future enhancement package.

Concerning the CTS Deploy Web Service, it is sufficient to configure and use this Web Service on the TDC system only. activation of customizing table logging (SCU3). Prerequisite for IMG Logging is: . InTechnical settings in ABAP Dictionary maintenance (transaction SE11), the function Log data changes has been selected for the respective customizing table or view. . The profile parameter rec/client is set for the appropriate clients. IMG projects ==> SPRO_ADMIN Transaction SCC1 copies changes from one client to another based on: . A task (don't need to release it) . A transport request . A transport request and its tasks ==> you can maintain a separate client for testing (unit test) changes to cross-client customizing objects, global settings and Repository objects require a Workbench transport request. SAP Software Change Registration (SSCR) ==> register developers, register SAP objects to be modified Customer name ranges can be reserved using the view V_TRESN, which enables developers to assign a specific name range to a package. The object directory is stored in table TADIR. This table is very central to your system's consistency. Object Directory lists: . All SAP standard Repository objects

. All customer-developed Repository objects The view V_TDEVC holds all packages in the SAP system, including all SAP packages. Release a transport Releasing a transportable transport request records a version of all the objects included in the transport request (that are included in the version management), and then exports the objects, that is, copies the objects from the database to an operating system file in the transport directory. Releasing a local transport request records a version of the objects in the transport request (that are included in the version management), but does not export them to a file at the operating system level. In the Version Overview, the active and modified versions are displayed in the development database, and versions saved as a result of released transport requests are displayed in the version database. Versions are generated when exporting a transport request. According to the tp parameter settings also when importing a transport request (not by default). A temporary version can be generated manually at any time by the developer. In the transport logs, the following return codes indicate the success or failure of the export: . 0: The export was successful. . 4: A warning was issued but all objects were transported successfully. . 8:Awarning was issued and at least one object could not be transported successfully. . 12 or higher: A critical error had occurred, generally not caused by the objects in the request. When a Repository object is transported to a downstream system, it exists there as a copy. Changes to copies, which are referred to as repairs, are made only in exceptional cases. In all customer systems, including the customer's development system, SAP-delivered Repository objects are copies. The original of those objects exists at SAP in the Development System. Repairs to SAP objects are referred to as modifications. development/correction, that is changes in the original system, and repair, that is changes in a system other than the original system. A list of all modified objects in the system can be shown with the Modification Browser (transaction SE95). Use Relocations of objects w/o package change if development work on objects is to take place in another SAP system temporarily. Special developments may be carried out in a separate SAP system, for example, so as not to interfere with the development process. This request type allows you to move the original location of objects to the target system. This function replaces "Transport of Originals" or C-type transports as of Release 3.x. Use Relocations of objects with package change when the development system of individual objects is to be changed on a permanent basis. This request type allows you to change the original location of objects to the target system and change the package of the objects at the same time. Due to the package being changed automatically, the objects have the desired transport attributes immediately after being imported into the target system of the request.

tp & R3trans (import process) The tp import all command does not process all steps for one request before moving on to the steps for the next request. For example, if you detect an error in an object that has already been released, you correct the object and release it again. 'Import all' then imports the whole buffer in the correct sequence and the faulty object is overwritten. Because the generation step is the last step and is performed only once for all objects in the requests, the faulty object does not affect your production system.

DDIC import & main import ==> R3trans | other steps ==> background jobs (RDD*-jobs started by RDDIMPDP) R3trans writes data using a standard transport format. Thus you can export data with an old R3trans version and import data with a new version of R3trans. You can also transport between different databases or operating systems. Note: Although exports and imports are independent of the R3trans version, the database platform, or the operating system, SAP does not support using tp or R3trans for transports between different SAP system releases. In addition to the steps performed by R3trans, tp triggers the so-called import dispatcher. The import dispatcher is a background job with the name RDDIMPDP in the SAP system, which performs steps in the import process. tp and RDDIMPDP communicate via table TRBAT. For every transport request, tp writes an entry in table TRBAT. To trigger the transport daemon RDDIMPDP in the SAP System, tp uses the operating system level tool sapevt. When RDDIMPDP is started, it checks the table TRBAT to find out if there is an action to be performed for the transport requests such as mass activation, distribution, or table conversion. It sets the header entry to R (for run from B=begin), and starts the appropriate RDD* program as a background task, reschedules itself, and exits. The activated program sets the status of the first entry in TRBAT to active (RC: 9999-->8888). Each of the required background tasks receives a job number generated by background processing. This job number and the step ID are inserted into table TRJOB by the RDD* jobs. The background tasks write their return codes in table TRBAT and delete the corresponding entry in table TRJOB. Return codes of 12 or less indicate that the step is finished. In TRBAT, the column TIMESTAMP contains the finishing time. When

all the necessary actions are performed for all transport requests, the header entry is set to F (for finished) by the RDD* jobs. tp monitors the entries in table TRBAT and table TRJOB. When the header entry in table TRBAT is set to F and table TRJOB is empty, tp copies the logs of the completed steps from directory tmp to log and deletes the corresponding entries in table TRBAT. If problems are detected by tp when monitoring table TRBAT and tableTRJOB, tp re-triggers RDDIMPDP through sapevt. RDDIMPDP automatically recognizes if a previous step is still active or was aborted by checking tables TRJOB and TRBAT. If a step was aborted, RDDIMPDP restarts this step. In client 000 and each client that is to receive imports, user DDIC must schedule the job RDDIMDP with event-based scheduling. Because import buffers and import queues are identical, use of tp and TMS produces consistent results. commands to start an import using tp at the operating system level: . tp import all <target SID> client=<client number> pf=TP_<domain name>.PFL Analogous to the default import of TMS, this command imports all waiting transport requests in the correct export sequence. Analogous to the TMS preliminary import, this command enables you to give individual transport requests priority. . tp import <transport request> <target SID> client=<client number> u0 pf=TP_<domain name>.PFL To ensure that the objects imported through an individual transport are not overwritten by an older version, only use unconditional mode 0, that is, the option u0, when starting individual imports. Using this mode is analogous to the TMS preliminary import and causes the transport request to remain in the list of requests to be imported. The tp unconditional import modes are as follows: . 0 Imported transport requests are not deleted from buffer. . 1 Ignore that the transport request has already been imported . 2 Overwrite originals . 6 Overwrite objects in unconfirmed repairs. . 8 Ignore restrictions based on table classification . 9 Ignore the fact that the system is locked for this type of transport Hint: The unconditional modes may vary in dependence of the SAP & tp release Import monitor ==> display tp status information from table TPSTAT To display all tp calls and errors for the current SAP system in the file SLOG, choose Goto tp System Log. To display the file ALOG containing the return codes of all individual transport steps for the current transport group, choose from the tp system log view Goto Transport Steps. To check if the transport requests in an import queue contain critical objects that should not be imported into the target system, go to the Import Overview (STMS Overview Imports ) and choose Queue Check Critical Objects. The object lists of the requests are checked to see if they contain critical transport objects. log files are named <source SID><action><6 digits>.<target SID>, where the action is represented by a single character and the 6 digits are taken from the corresponding transport request. Each log file can be seen in the SAP system. You can choose different detail-levels in the log file.

All logs in the transport environment are stored in the transport subdirectory log. These logs include logs created by tp (ULOG, SLOG and ALOG), and logs created by the various transport tools. Examples of log files that may be written during the import process of request DEVK900132 include: . DEVH900132.QAS - import log file . DEVA900132.QAS - dictionary activation log file . DEVI900132.QAS - main import log file . DEVM900132.QAS - enqueue log file . DEVV900132.QAS - versioning log file . DEVR900132.QAS - user-defined activities . DEVG900132.QAS - generation log file The accompanying text is more significant than the value of a return code. To display the text of a specific tp return code, use the tp command tp explain rc <value of return code>. freezing landscape With the following tools an administrator can force code freezing for a system landscape: . stop exports by creating a file in transport directory: /usr/sap/trans/bin file: T_OFF.<SID> The first line of this file will be displayed, whenever you start an export from the Transport Organizer. . stop imports by creating a file in transport directory: /usr/sap/trans/tmp file: NOIMPORT.<SID> The content of this file is not evaluated. CUA:

If the central system and the child system are in different SAP systems, three RFC connections are required. The RFC destination names must be identical to the logical system you want to address with the RFC connection. The RFC connection must be specified in uppercase letters. You can also use trusted RFC connections for the CUA communication. This increases the security of your CUA communication... For each field of transaction SU01, you can use transaction SCUM to determine the system in which the administration of the field content can be performed. If you are using CUA, you can only create users in child systems using transaction SU01 in the central system. The creation of users is possible using the Systems, Roles, and Profiles tab pages. directory servers (LDAP & SM59 tcodes) The SAP system must be able to log onto the directory service for the various LDAP operations. The directory services themselves use various authorization concepts (from anyone being able to retrieve information, to extensive identity checks using a directory server). To do this, enter the logon data for an LDAP system user in the System Users area in transaction LDAP using a specific user ID. In addition to the name and credentials (password entry), this also includes information about whether a user may access the directory service to read only, and where the credentials are stored. Note: The name of the created user ID corresponds neither to a user in the SAP system, nor to a user in the directory server. Rather, it is a key that

is required for the connection data. The LDAP system user is stored in Customizing for the directory service connection and has nothing to do with the SAP users maintained in transaction SU01. SAP has decided not to develop its own schema, but rather to support the schemata available on the market. As the delivered schemata for the various products do not know many SAP data fields, you must use schema extensions to make these fields known. An extension of this type is provided and imported into the directory server in a special format (LDIF for .LDAP Data Interchange Format.). You can easily create files of this type in the SAP system for all certified and some additional products using report RSLDAPSCHEMAEXT. Synchronyzation ==> report RSLDAPSYNC_USER Log for all LDAP operations ==> LDAPLOG or SLG1 p. 551 tadm12_1 TADM51 combination of oracle processes and memory buffers = oracle instance (oracle listener is not part of a instance) rac (n istances for 1 db): global --> datafiles, controlfiles, db profile; local --> online redo logs, archive logs, alert log change SAP<SCHEMA-ID> oracle user password only using brconnect ... If you perform a full backup, after backing up all data in the database, an additional piece of information (the catalog information) is written to the control file by the Recovery Manager, which makes it possible to subsequently create incremental backups. A whole backup creates a backup of all data without the catalog information. In terms of the database data, there is no difference between the whole backup and full backup. an incremental backup can only be created when RMAN is used as the program for writing data to backup media. To perform a complete database reset, you need Complete offline backup | Consistent online backup When performing an incremental backup, an extra database block verification does not need to be performed. While having multiple components in one database requires the MCOD tablespace layout, having one database per SAP system does not necessarily require the MCOD tablespace layout. the number of data files should be about 100 when the database has reached its expected size.

All tablespaces created by SAP installation tools in the "old" tablespace layout were created as dictionary-managed tablespaces. SAP recommends using Locally Managed Tablespaces. These are the default tablespaces if the new MCOD tablespace layout is created by the installation tool. To switch from dictionary-managed to locally managed tablespaces on existing databases, a reorganization is necessary. Whether this reorganization can be done online or must be done with export/import depends on the database version. With locally managed tablespaces, each data file has a bitmap listing used and free blocks within the data file. extent size will no longer be determined by storage parameters. The extent size is either identical for all extents (UNIFORM) or automatically chosen depending on the segment size (AUTOALLOCATE). The kind of extent allocation (UNIFORM or AUTOALLOCATE) is determined at the creation of the tablespace and cannot be changed. SAP uses locally managed tablespace with automatic extent allocation as of Oracle 9.20. Exceptions: tablespace PSAPTEMP is created locally managed, but with a UNIFORM extent size. ablespace PSAPROLL is created as dictionary-managed tablespace. For rollback, a new PSAPUNDO tablespace should replace PSAPROLL. In SAP BW systems, tablespaces holding fact tables or aggregates should be created with a uniform extent size of 1 MB. To provide transactional consistency and read consistency, Oracle uses rollback segments in a rollback tablespace or, as of Oracle 9i, undo segments in an undo tablespace. When a database transaction starts, Oracle selects one existing rollback or undo segment for this transaction. Whenever table data is changed, the data is changed in its original block, after the old data was copied into the assigned rollback or undo segment. Whenever a rollback is necessary (for example, a transaction was canceled or an instance recovery is performed after an instance crash), the original data will be recovered from the rollback or undo segments (transactional consistency). Whenever a process reads data that is actually changed by another process, the reading process reads from the rollback or undo segments (read consistency). undo_retention This is the time, in seconds, that Oracle keeps undo data (called undo retention time). After this time period, undo data can be overwritten. brconnect (brconnect -u / -c -f check.) DBCHECKORA DB17 In addition to the output in the log file, the database system check also writes warnings and errors into the database table DBMSGORA. This table can be viewed with transaction DB16 To make sure that statistics are up to date, SAP recommends that you run brconnect -f stats -t all regularly. Run brconnect -f stats -t missing whenever new tables are created that do not have statistics. It can be scheduled with transaction DB20 (choose Global statistics Create missing). Such tables will be reported by the database system check. Missing statistics will be also collected in the standard run brconnect -f stats -t all For optimal monitoring of the Oracle database within the CCMS alert monitor,

you must schedule the database system check to run daily you cannot create dictionary managed tablespaces with BRSPACE due to its disadvantages. BRSPACE always creates locally managed tablespaces resize data files: brspace -f dfalter and select Resize data file If you need to create a data file on a non-standard sapdata disk, therefore you should create the directory path <drive:>\oracle\<SID>\sapdata<n> on Windows before and create the new data file there: you can use brspace -f dfmove To move or rename a data file, the database needs to be shut down. BRSPACE will normally only shut down the database, if SAP is not running. In databases having multiple components in one database (MCOD) the tablespaces SYSTEM, PSAPROLL/PSAPUNDO and PSAPTEMP must not be set offline, because these tablespaces are used by all schema users. reorganization had to be performed offline and took a considerable amount of time when large tables or whole tablespaces had to be reorganized. The tool BRSPACE supports online table redefinition as well as traditional reorganization methods. Export tables and Import tables (brtools, segment management) Use these functions to reorganize tables using the old reorganization method by exporting tables, dropping them, and importing them again. Normally, tables are reorganized using the BRSPACE function Reorganize tables, which will perform an online redefinition of tables. But online redefinition is not possible in Oracle 9i for tables containing LONG fields (existing, for example, in SAP table clusters). As of Oracle 10g, online redefinition will be possible for tables containing LONG fields; LONG fields will then be redefined as LOB (Large Object) fields. Consequently, reorganization tables containing LONG fields in Oracle 9i must be performed by export/import. BRSPACE supports online table definition only in locally managed tablespaces Data Pump export and import utilities should be performed as a DBA database user. To do this, use the option "-u" when you call BRSPACE, for example: brspace -u system/<password> -f tbexport -l expdp To transform dictionary-managed into locally managed tablespaces with Oracle 9i, perform a normal online reorganization of all tables within a tablespace into a new tablespace, with the following settings and additional steps: - Before the reorganization, create one new locally managed tablespace storing tables and indexes (contents: both). - Start BRGUI or BRTOOLS and choose Segment management Reorganize tables. Specify the name of the tablespaces to be reorganized, and enter the "*" character in the table field. No other options need to be changed. Segment shrinking offers an alternative to reorganization for defragmenting a segment and gaining more free space. Segment Shrinking is available as of Oracle Release 10g

Compared to an offline reorganization, there is no downtime. While an online reorganization temporarily requires twice the space, segment shrinking does not require any additional space requirements. Even if segment shrinking is mainly an online operation, table locks (TM enqueues) may occur. The best solution is to run segment shrinking only at times when no long-running transactions are active, and change the underlying table. rebuild indexes: In the menu of BRTOOLS, you can enter a list of tables. If you enter tables instead of indexes here, all indexes of the specified tables will be rebuilt. Archiver Stuck problem Create a dummy file of about 10 times the size of an online redo log file in the archive directory with the name_delete_me_on_archiver_stuck. An archiver stuck cannot be avoided completely; when it occurs, you simply delete the dummy file, giving you some time to run BRARCHIVE to save and delete old archived redo logs. Change the log_archive_dest parameter: Only change the path; the last part of the original log_archive_dest (see above) is not a directory, but the prefix for the names of the archived redo logs. Use the command brspace -c force -f dbparam -p log_archive_dest -v <new drive>:\oracle\<SID>\oraarch\<SID>arch on Windows and brspace -c force -f dbparam -p log_archive_dest -v /oracle/<SID>oraarch/<SID>arch on UNIX. To check if everything is okay, perform a manual log switch with brspace -c force -f dbalter -a switchlog and check if a new archived redo log file was created in the new orarch directory. BRARCHIVE 7.00 is able to find and save offline redo log files after changing the archive directory, which offers the following simple procedure for resolving an Archiver Stuck problem: . Use BRSPACE to change the value of the Oracle parameter log_archive_dest[_1|_2] so that it points to a new archive directory with enough free space. . Start BRARCHIVE to save and delete offline redo log files. Make sure that the offline redo logs can be saved twice on different media. To do this, run brarchive -cds twice. . Then change the value of the parameter log_archive_dest back to the original value. To check if a tablespace is currently in backup mode, you can use the faster (and cheaper) command brspace -c force -f dbshow -t tslist. Check the Back column to see if any tablespace is in backup mode. Sometimes it happens that the whole database or server crashes during an online backup, for example when a reboot is executed while BRBACKUP is running. In this case, a normal startup is not possible ==> The problem can also be solved without performing a recovery, which can take a long time depending on the database activity and the duration of the

tablespace being in backup mode. To avoid a long recovery, use SQL*Plus and startup the database into mount state. Then execute the command ALTER DATABASE END BACKUP and open the database. Oracle architecture Every running Oracle database is associated with an Oracle instance. Moreover, every Oracle database needs its own instance.Using real application clusters (RAC), one database is served by two or more instances. Every time an Oracle instance is started, a shared memory region called the System Global Area (SGA) is allocated. SGA ==> buffer cache (DB_CACHE_SIZE) + shared pool (SHARED_POOL_SIZE) + redo buffer (LOG_BUFFER) + large pool + ... DB_CACHE_SIZE: Setting this parameter activates the dynamic SGA. It defines the size of the buffer cache and thereby replaces the previous parameter, DB_BLOCK_BUFFERS, which has become obsolete. DB_BLOCK_BUFFERS cannot be used simultaneously with SGA_MAX_SIZE or DB_CACHE_SIZE. The following cannot be changed dynamically or are only changeable by restarting the instance: . Maximum SGA size: SGA_MAX_SIZE . Redo log buffer: LOG_BUFFER . Java pool: JAVA_POOL_SIZE SHARED_POOL_SIZE: The shared pool (shared memory area in the SGA) is used by Oracle to hold several key memory structures. Most important among these are the data dictionary cache and the shared SQL area. main components of the shared pool are: . The shared SQL area, where parsed SQL statements are cached for shared access to all shadow processes ==> The ratio of reloads to pins should be, at maximum, 0.04. | The pin ratio should be larger then or equal to 95%. . The Data Dictionary cache (row cache), which holds the Oracle Data Dictionary information, including cost-based optimizer statistics ==> The ratio of user calls to recursive calls should be at least 2 or higher. | The Data Dictionary cache quality should also be greater than 80%. User calls: A shadow process accessing the shared SQL Area for parsed SQL statements Recursive call: The Data Dictionary cache making a physical read to load Oracle Data Dictionary objects from the system tablespace Automatic Shared Memory Management (ASMM) As of Oracle 10g, parameter SGA_TARGET is also available. You can set this instead of parameters DB_CACHE_SIZE, SHARED_POOL_SIZE, JAVA_POOL_SIZE, LARGE_POOL_SIZE, and STREAMS_POOL_SIZE. If necessary, Oracle configures and then adjusts the individual SGA components within SGA_TARGET. SAP recommends that you use SGA_TARGET (if you use it at all) if DB_CACHE_SIZE and SHARED_POOL_SIZE are set to sufficiently large values. In connection with SGA_TARGET, these parameters specify lower limits for the area values, below which the values cannot fall. Program Global Area (PGA): The larger you set the *_AREA_SIZE parameter or PGA_AGGREGATE_TARGET, the more PGA operations can be performed in memory.

Administrators of SAP NetWeaver BI databases should pay attention to the tuning of the Program Global Area due to memory-intensive SQL operations typical to the data warehouse area (large sorts and joins). The automatic Program Global Area is particularly suited to and beneficial when using with SAP NetWeaver BI applications. However, it can also be beneficial when used with other non-BI SAP applications. The database administrator now only has to determine the maximum total PGA memory of an instance (PGA_AGGREGATE_TARGET). To maximize overall performance through efficient memory usage, the Oracle server uses an intelligent algorithm to divide this memory between the existing Oracle server processes depending on demand. automatic PGA is activated if - WORKAREA_SIZE_POLICY = AUTO Using this parameter, you can set the automatic PGA tuning dynamically to AUTO or switched off (MANUAL). Oracle Release 10g . Default = AUTO | Oracle Release 9i . Default = MANUAL - PGA_AGGREGATE_TARGET = a value dynamically adjustable OLTP: PGA_AGGREGATE_TARGET = <Total physical memory> * 20% BW: PGA_AGGREGATE_TARGET = <Total physical memory> * 40% obviously: SGA_MAX_SIZE + PGA_AGGREGATE_TARGET <= <phys. memory of DB server> data buffer quality is based on the ratio of physical reads to the total number of reads. The lower the ratio, the better the buffer quality. The data buffer quality should be better than 94,0% but you also consider the ratio of buffer gets to user calls. In a well-tuned system, this ratio is approximately 15 (if reads/user calls >> 15 ==> real data buffer quality not good also if > 94%). The Cache advisory statistics (SGA Monitor sub-monitor is called as detailed analyses via Performance Statistical Information SGA Monitor.) screen contains rows that predict the number of physical reads for the cache size corresponding to each row. In well-tuned systems, the wait events make up about 60% of the response time. Otherwise, the proportion may be much higher, which has a negative effect on the response times. Wait event tuning can therefore frequently bring about a significant improvement in database performance. When Oracle needs to access a data block and the block is already in the buffer cache, a logical read occurs with no physical read. The process is able to read the block without the occurrence of any wait events. Large amounts of CPU time could be consumed on significant numbers of logical reads, and the wait event interface will have nothing to show for the elapsed time. Total wait time describes the sum of idle wait time and busy wait time. Database response time consists of the wait time and the CPU time. SAP recommends a CPU idle proportion of at least 30% per hour. Which of the following wait events are busy wait situations? Answer: B, C Only B (log file sync) and C (db file sequential read) are busy wait situations. SQL*Net message from client is an idle wait event and the CPU time is not covered by the Oracle Wait Interface. 2. Which of the following wait events may be caused by full table scans? Answer: C db file scattered read may be caused by full table scans.

AWR (enables historical performance analysis) As of Oracle 10g, Oracle provides a multitude of new features that facilitate workload analysis, space requirements, and database performance. The database creates a number of new classes of statistics and provides a repository to store these statistics permanently for later analysis. To fulfill these requirements, the Automatic Workload Repository (AWR) was implemented. The data that is collected within the functions that are described here is stored in the SYSAUX tablespace. AWR snapshots are taken automatically by the database instance from a background process called MMON. AWR snapshots are gathered by default once per hour and they can also be triggered manually, but this is usually not necessary. The length of time in which AWR keeps data around before purging is referred to as the snapshot retention period. By default, the retention period is set to seven days. The retention period and the snapshot interval can be changed by the administrator. Automatic Database Diagnostic Monitor (ADDM): Tuning recommendations based on AWR: ADDM provides system performance analysis based on the AWR data and makes appropriate recommendations for corrective actions. ADDM automatically detects and diagnoses common performance problems: Hardware issues related to excessive I/O | CPU bottlenecks | Connection management issues | Concurrency issues, such as contention for locks ... Workload Reporting monitor in the DBA Cockpit lets you display very detailed reports from the Oracle Active Workload Repository. To access this monitor, choose Performance Wait Event Analysis Workload Reporting in the DBA Cockpit. ASH=Active Session History contains information about the active session of the database instance. Expensive statements are defined as all SQL statements that cause the database to read many blocks (from disk or buffer). Usually statements are called expensive if the database reads more than 5% of the total blocks it reads for the answering of those statements ST06 - transaction profile: Sort the Total Database Time column in ascending order and sum it up. Transactions that are causing more than 5% of the summed (total) database time might be worth tuning. ST04 Call ST04 and choose Performance SQL Statement Analysis Shared Cursor Cache: Analysis of Shared Cursor Cache. For Buffer gets, choose a number equal to 5% of logical reads from the ST04 entry screen. There might be expensive statements causing less than 5% load on the database, but these usually have little impact on system performance, even when optimized. Logical reads: Number of Oracle buffer blocks read for the statement from the data buffer. All logical reads from the Oracle database are in reads from the data buffer. To provide the data in the data buffer a physical read must sometimes be executed. Buffer gets: Number of Oracle buffer blocks read for the statement from the data buffer. Identical in meaning to logical reads. Physical reads/disk reads: Number of Oracle blocks read for the statement from the hard disk search for full table scans using the following method: . Enter transaction ST04. . Choose Performance Additional Functions GV$-Views. . Double-click on the line V$SQL_PLAN under ..V$Views header (the

right-hand column). . Right-click on the Operation column header and choose Set Filter. . Select for the operation TABLE ACCESS and choose Enter. . Repeat the previous step for the Options column with the value Full. . Right-click the OBJECT_NAME column header and choose Sort in descend order. The result will be a list of tables accessed by full table scans, sorted with Z* tables on the top. These table accesses might be tuned. Oracle locks: TX (transaction enqueue): parallel change of a data record not possible TM (DML enqueue): a complete object has to be protected against changes (for example, as part of an index rebuild or a consistency check using VALIDATE STRUCTURE). ST (space transaction enqueue): This enqueue is held in dictionary-managed tablespaces within extent allocations and releases. The Oracle Lock Monitor can be called by using the DBA Cockpit (transaction ST04) and choosing Performance Wait Event Analysis Lock Monitor. If exclusive lock waits occur, both blocking and blocked processes are displayed. For each blocking situation, a tree of waiting processes is shown. Column PID shows the process IDs of the host processes. Usually, this is the process ID of the SAP work process. indexes (B* tree, bitmap) The advantage of using bitmap indexes is greatest for low-cardinality columns, that is, columns in which the number of distinct values is small compared to the number of rows in the table. Bitmap indexing benefits data warehousing applications that have large amounts of data and ad hoc queries, but a low level of concurrent transactions. Bitmap indexes are not suitable for OLTP applications with large numbers of concurrent transactions modifying the data. These indexes are primarily intended for decision support in data warehousing applications where users typically query the data rather than update it. The indexes normally set up by the SAP system are B* tree indexes. Bitmap indexes are particularly significant in the SAP NetWeaver BI system (for example, on the characteristic columns of InfoCubes). For queries where the primary index (which consists of the key fields) cannot be used to determine the results, for example, because none of the primary index fields were specified in the WHERE clause, the RDBMS scans through the entire table. To avoid this, you can define secondary indexes to significantly reduce the number of data records searched. Secondary indexes can also be unique indexes. Secondary indexes creation: As a rule of thumb, an index should contain no more than four fields. Position selective fields to the left in an index. As a rule of thumb, you should not create more than five indexes for each table. Do not change an index created by SAP. Sometimes the optimal index combination can be found only by trial and error. Creating or changing an index can take from several minutes to several hours and blocks the entire table. This causes serious performance problems during production operation. Transaction DB05 displays the selectivity of a field combination. Use this information

to determine if an index should be created. The number of distinct values shows you how many different combinations of fields exist. This number should be close to the number of table rows for the combination of all of the key fields of the index. In general, an index it is not effective if it returns many rows.

The process of activating an index for a large table in the database is especially time-consuming. During this process, INSERT, UPDATE, and DELETE operations affecting the corresponding table are blocked. Therefore, avoid creating new indexes for a large table during business hours. After activating a new index, you may need to generate new table access statistics so the optimizer can consider the new index when calculating the execution plan. database statistics The cost based optimizer chooses the access path with the fewest expected I/O operations. The optimizer's decision is based on estimates. Whether the estimates are correct depends on the quality of the database statistics. Statistics are not required in the following cases if you are using Oracle databases <= 9i: using the rule-based optimizer or if an explicit RULE hint is specified | SAP table pools and table clusters | exceptions according to DBSTATC | Oracle DDIC objects | INSERT statements Database statistics are required, in general, as of Oracle 10g (cost-based optimizer) There are different types of statistics available: table, index, and column statistics. Histogram information, another type of database statistics, is stored in the Oracle view DBA_TAB_HISTOGRAMS. Histograms simply describe the distribution of column values between the lowest existing value (LOW_VALUE) and the highest existing value (HIGH_VALUE). Histograms are optional and represent more detailed column statistics. With no histograms available, the Oracle database assumes an equal distribution of the values inside a column. However, Oracle database histograms are not generally recommended in SAP environments. Creating statistics does not block any database objects. In theory, therefore, you can create statistics while the system is running. However, since the creation of statistics involves an additional CPU and I/O load, it should be carried out during a period of minimal workload if at all possible.

The creation of a new statistics is based on a 50% criterion: New table statistics will be created if the table content, that is, the number of rows, has changed at least by 50%. Both an increase or a decrease of table content is taken into account. As of Oracle 10g, Oracle database statistics can also manage completely automatically. This is implemented in the database as the DBMS_SCHEDULER job GATHER_STATS_JOB. However, this method is not yet supported in the SAP environment, and is therefore deactivated to avoid conflicts with the BRCONNECT method. In new SAP installations as of SAP NetWeaver 7.00, this automatic function is deactivated by SAPINST. For mass analysis and mass defragmentation of nonpartitioned B*Tree indexes, you can use report RSORAISQN. This report provides different heuristic algorithms for fast and exact analysis and provides different methods for defragmentation, such as coalesce or rebuild online. It is restartable and controlled stoppable. Different possibilities of parallelism and history of storage qualities are offered by this report. Furthermore, it gives you the ability to monitor currently running operations and to estimate further runtime To identify I/O contention in the database, use the Filesystem requests sub-monitor in the DBA Cockpit (performance-wait event analysis). This monitor helps you to minimize the time needed to read or write data from or to a file, so that you can identify the frequently used data files and put them on separate disks to avoid contention. Data file activity has an important effect on database performance. Tuning the PGA over-allocation count (from PGA monitor, V$PGASTAT view) This value specifies how often the value of PGA_AGGREGATE_TARGET did not meet the minimum memory requests of the server process. As a result, Oracle had to add additional PGA memory internally again over and above the maximum value specified in PGA_AGGREGATE_TARGET. If PGA_AGGREGATE_TARGET is set optimally, the value of over allocation count should be zero. If over allocation count is greater than zero, the value of PGA_AGGREGATE_TARGET is too low. You must then increase PGA_AGGREGATE_TARGET. cache hit percentage This value shows the performance/efficiency of the Oracle automatic PGA memory management since the instance started. The ideal result of 100%, which is not usually achieved, means that the work areas of the server processes were the optimum size for all queries and that no one-pass or mult-pass operations occurred. to tuning you need to: . Avoid PGA memory over-allocation Set PGA_AGGREGATE_TARGET at least large enough so that the estimated over-allocation count = 0. If you set PGA_AGGREGATE_TARGET smaller (that is, too small), Oracle will automatically over-allocate this minimum amount of memory. . Maximize the PGA cache hit percentage

The optimum, ideal cache-hit-percentage value would be 100%; however, this is only achievable with a very high memory requirement. Here, you should assess the marginal benefit of each additional MB and, based on that, decide whether it justifies the memory usage. The optimum value for PGA_AGGREGATE_TARGET is the value at which a further increase would not result in any relevant increase of the PGA cache hit percentage. A cache hit ratio of 90% for BW systems and of 60% for OLTP should be sufficient ---|---|--Abap process flow:

A SAPWeb Dispatcher can distribute requests for only one SAP system. If multiple SAP systems are required, you have to set up and start separate SAP Web Dispatcher processes for each of the respective systems (which can run together on one computer). Instead of maintaining the same threshold values in every system for CCMS monitor infrastructure, we recommend that you maintain the values in the central monitoring system and then distribute them to the monitored SAP systems using the transport system. The prerequisite for transporting the threshold values to other SAP systems is that you have stored them in properties variants (Containers in which system monitoring settings can be saved).
*) C_TADM51_70 WA.pdf Bapis are specialized remote functions calls tadm51 v. anche Sample Questions Unit 4 p. 34

a,d - b,d - true - b,c tadm10_2 p. 494 (rivedere solution 3 ... da p.94) Agate can reduce volume data transferred to web browser: compression with gzip | ICM is a separate process, icman at operating system level | ISC: internet server cache (important part of ICM) ==> icm/HTTP/server_cache* | icmon at operating system level to monitor ICM and generate requests (workload simulation) | an HTTP request handler is an abap class that implements interface IF_HTTP_EXTENSION, this interface contains HANDLE-REQUEST method which is called by ICF ... | SICF, Edit-Recorder / SICFRECORDER (db registration/analysis of HTTP requests) !! | itsp/enable ==> 1, em/global_area_MB increased for integrated ITS | RSTRC000 report for tracing | sap web dispatcher (reverse proxy & load balancer): needs to know the port at which it is to receive HTTP requests (icm/server_port_<x>) and on which host (rdisp/mshost) and with which HTTP port (ms/http_port) it can access message server | you can monitor web dispatcher with icmon or via /sap/wdisp/admin URL ... | stateless (each request is independent of all other requests) & statefull web apps (transactional concept, user session stored in the appl server) | ABAP authorizations: all authorizations are permissions, everything that is not explicitly allowed is forbidden | SUIM: user information system | ALE: application link enabling | UME persistence manager supports data partitioning (attribute-based, user-based, type-based): user data for different user types can be stored in different data sources | principle: user, user account, group, role | with abap system data source for UME ==> abap roles <--> UME groups | java users types: standard, technical users, internal service users | j2ee security role (only in VA), UME role; permissions, actions (collection of permissions), roles | IDoc=intermediate documents | RFC: based on CPIC & TCP/IP | BAPIs represents methods (normally RFC-enabled function modules) for business objects (BO) in an Sap sys |Web services: WSDL, UDDI, SOAP | Enterprise SOA: implemented across systems, abap or java, no own database, new applications without modification of underlying system, high flexibility | SLD: content based on CIM (common information model) standard (http://www.sdn.sap.com/irj/sdn/nw-sld) | standard ccms nodes (monitors) only collect local data | short dumps (ST22) are stored in the system for 14 days | suggestion for notes search: search for "composite notes" for the topic area ... | monitoring AS java ==> JMX (Java mangement extension standard) monitors, CCMS agent SAPCCMSR for data collected by CEN (system for central monitoring), nwa: monitoring browser, VA: monitoring service | Logs from java AS are formatted in a special way (ListLog) for use in Log Viewer | logging: recording normal and exceptional events, tracing: recording the process flow of an application tadm10_1 (rivedere sezione ssl on NW AS Java) composite applications = xApps | J2EE 1.3 specification | java enterprise runtime, j2ee components, application layer | a sapstartsrv process is started for each Sap instance and does not terminate even if the instance is stopped, reads start profile, 5$$13 port (SAPMC) | java stack is started & stopped by the java startup and control framework = jcontrol & jlaunch | -Xmx max heap size, -Xms initial heap size, XX:newsize initial young generation, XX:maxnewsize max young generation, XX:maxpermsize max permanent generation, ... | offline configuration editor: inappropriate use can destroy AS java | telnet: man, lsc, jump <node id>, ... | you add an additional server process to a java instance with the config tool | dev_ms contains HW key | customer developments, customer enhancements, modifications | data element --> semantic data of a field, domain --> technical data of a field | a physical runtime system does not have to exist in the background for each of the 4 stages of the java development landscape test di esempio: http://it.toolbox.com/blogs/sap-library/sap-on-oracle-certification-preparation-c_tadm51_70-testyour-basis-skill-33751 - Packages are created and managed with ==> Package Builder (SPAK) - Users are assigned authorizations using roles http://it.toolbox.com/blogs/sap-library/sap-on-oracle-certification-preparation-c_tadm51_70-testyour-oracle-skill-33079

http://it.toolbox.com/blogs/sap-library/sap-on-oracle-certification-preparation-c_tadm51_70-testyour-oracle-skill-part-2-33750 certificazione ==> http://www.sap.com/services/education/certification/certificationtest.epx? context=[[C_TADM51_70|G]]|