1 2 3 4 Previous Next

SAP on IBM i

48 Posts

In this update for SAP on IBM i, we would like to call your attention to an issue that we have recently discovered in the SAP kernel releases 7.40 and 7.41. Note that this only applies to you if you have configured ES/TABLE=SHM_SEGS. If you are running with the default ES/TABLE=UNIX_STD, you will not experience this problem. While ES/TABLE=UNIX_STD is the default and used to be significantly faster in IBM i releases prior to IBM i 6.1, the alternative ES/TABLE=SHM_SEGS can be significantly faster in IBM i releases 6.1 and higher when hardware with a large number of processors is used.

If you follow our blog, you may remember that we introduced the memory model ES/TABLE=SHM_SEGS for SAP on IBM i through the update in July 2013: SAP Memory Model on IBM i. At that time, the switch was enabled and tested for SAP kernel release 7.20 at patch level 220.

With SAP kernel release 7.40, SAP has introduced a new paradigm for memory management in SAP. Prior to 7.40, a large set of memory management related profile parameters had hard-coded default values, while with 7.40 and beyond, these default values will be calculated based on the size of the main storage available (PHYS_MEMSIZE). The details of these changes are documented in SAP Note 1864189.

Even though not listed in this SAP Note, profile parameter ES/SHM_MAX_SHARED_SEGS will also be calculated in 7.40 and show values between 1 and 9 depending on the available main storage. When ES/TABLE=SHM_SEGS is configured and ES/SHM_MAX_SHARED_SEGS is calculated to a value larger than 1, your SAP instance will not come up because the dispatcher discovers a conflict between the parameters ES/SHM_MAX_SHARED_SEGS, ES/SHM_MAX_PRIV_SEGS and ES/SHM_PROC_SEG_COUNT. We are going to provide a kernel patch for release 7.40 and 7.41 in the near future. Until the patch becomes available, you can follow the instructions of SAP Note 1998233 to work around that problem.

We would like to inform you about the following offerings by IBM for the SAP on IBM i community.  They will take place in the United States and Germany, but the presentations will be given in English language and are open to anyone around the world who is interested.


SAP on IBM i Summit Customer Event (May 15th-16th in Rochester, MN)

The first ever SAP on IBM i Summit customer event will be held on May 15th and 16th in Rochester, MN.  The event is exclusively focused on those customers that run SAP workloads on IBM and will include topics such as:


  • IBM i 7.2 and POWER8 announcements
  • IBM i and DB2 for i enhancements for SAP
  • SAP strategy and roadmap for IBM i
  • Leveraging advanced DB2 for i technology for SAP workloads
  • Eliminating downtime for SAP Software Update Manager functions
  • DB2 for i optimization and administration with SAP
  • Integration of PowerHA into SAP resiliency solutions
  • Leveraging the SAP on IBM i Center of Excellence

Speakers include SAP on IBM i experts, including developers and experts from the IBM i Development Lab in Rochester and from the SAP on IBM i Porting Team in Germany.

Business partners and IBMers are welcome to attend along with their clients.  There is no cost for attending the event.

For more information and to register, please go to the following website: http://www.ibm.com/partnerworld/page/stg_com_sys_ibmi-sap-summit.

SAP on IBM i Database Education (April 1st-4th and April 8th-11th in Rochester, MN)

Two upcoming SAP on IBM i classes have been scheduled that you might be interested in.  Both classes will be held in Rochester, MN.

The first class is scheduled for April 1st to 4th and is SAP NetWeaver on IBM i: Database Administration.  This is exact same class as SAP's ADM525 course and is part of SAP's NetWeaver administrator curriculum.  The IBM course code is DB2iSAPADM.

The second class is scheduled for April 8th to 11th and is SAP on IBM i DB2 Optimization.  This course focuses on optimizing and tuning DB2 for IBM i for SAP workloads.  The course code is SQLOPTSAPoni.

You can register for both workshops at http://www.ibm.com/systems/power/software/i/support/workshops/enrollment/form.html (select the workshop from the drop-down list on top of the screen under “Workshop summary” and then select “Enroll Now”).

DB2 – SAP on i Technical Forum (April 28th-30th in Frankfurt, Germany)

This technical education class for customers and IBM Business Partners is primarily focused on SAP on i System Administrators and Basis Consultants and will explain how to position and optimize the performance of an SAP on i implementation in comparison to other databases.  It will include topics such as:


  • Overview of generic database storage, access and processing problems and solutions
  • Overview of specific database management systems
  • Why the entire stack matters - or: Why moving to a new DBMS means moving to a new stack
  • Best Practices for SAP on i
  • Best Practices for DB2 for
  • iTools, techniques and next steps for proactive and reactive analysis and tuning

This class has a more technical focus and is different to the Rochester Summit Event (May 15th to 16th) which
provides a strategic positioning and roadmap for SAP on i customers.

Speakers are developers and experts from the IBM i Development Lab in Rochester.


Attendance fee: 2,500 €.

For more information and to register, please go to the following website: http://www.cancom.de/cn-713--ci-712--DE-WFS-de_DE-EUR.html?aid=754. Please note that the website is only available in German, even though the presentations are given in English. If you do not speak German but need more information or want to register, please send an e-mail to kristina.diesem@cancom.de.

Welcome in SAP on UBM i update! This time we will describe how to use ssh to do simple administrative tasks.

There are situations when one wants to access the SAP system on IBM i in the same way as a SAP system on UNIX, e.g. someone without IBM i knowledge needs to test the effect of modifications of SAP-Profiles.


SAP-Note 1966316 describes the steps necessary to set up a UNIX-like Environment for <SID>adm. This updates the paseinlpgm and the startsap/stopsap programs and creates/updates some UNIX-profiles in /home/<sid>adm.


You need an SSH-Client for your Windows-PC in order to use these features. A popular one is PuTTY, a free Telnet/SSH client maintained primarily by Simon Tatham.

Then log on to your IBM i System as user <sid>adm: After starting PuTTY and entering the host name you are presented with the 'login as:'-Prompt where you type in the user-name, hit <ENTER> and type in the password and hit <ENTER>.


Then you are at the mercy of the minimalistic '%'-Unix prompt and may start working --- but only if you know what you are doing.


To stop the SAP-System use the 'stopsap' command and to start it the 'startsap' command.


To leave the SSH-Session type 'exit'.

Customizing the UNIX-Environment:


The file .login_shell in the home-directory <sid>adm determines which UNIX-Shell is used for login. Standard shells are the preset C-Shell (/QOpenSys/bin/csh) or the Korn-Shell (/QOpenSys/bin/ksh93).


You can customize these shells by editing the .login (for csh) or the .profile (for ksh93). See the appropriate Unix documentation on how to do this.

Welcome at the SAP on IBM i update! This time we will focus on the new features of the CONVUSRCPT tool:


The current situation

All SAP Kernels in maintainance support the new userconcept

System installation and system copy tools (SWPM, Software Provisioning Manager) always create new SAP systems and instances with new userconcept

System upgrade and system update tools (SUM, Software Update Manager) require start releases >=7.0 to be on the new userconcept

After upgrading from start releases <7.0 an immediate switch to the SAP Kernel 7.20 and the new userconcept is strongly recommended

With the first SAP Product release that will not support an upgrade path from start releases <7.0 we will technically disable the classic userconcept


So there is no way to circumvent the conversion from the classic to the new userconcept!

...but we know the obstacle

In collaboration with customers and consultants we learned about two constantly growing directory classes.

The transport directories

The j2ee directories below the instance directories


Having at least one of these directories populated with many files slows down the conversion and drastically increases the downtime.

From this finding we invented a new conversion method ADVANCED consisting of three steps with two uptime and one shorter downtime step, while the former method now called STANDARD consists of one single large downtime step.


The three new steps creating the ADVANCED method are:

PRE_ADV  (uptime)       Conversion of the transport directories and preparation of the j2ee instance directory trees (if existing)

ADVANCED (downtime) Conversion of the remaining part (not transport or j2ee directories, but with the kernel lib and user profiles)

POST_ADV (uptime)      Conversion  of  the j2ee instance directory trees (if existing)


During our inhouse tests and the pilot phase on customer side we made the experience using the new ADVANCED method can shorten the remaining downtime step ADVANCED in comparison to the former conversion (now called STANDARD conversion ) by a factor of 5-10.


For more information, please see the rewritten SAP Note 1149318 'User concept conversion using the tool CONVUSRCPT' .


Stay tuned...

As the development team intends to maintain a consistent format, editions of the series will continue to be published.

On November 15th 2013, IBM i 7.1 technology refresh 7 became generally available. It can be ordered from IBM as PTF group SF99707. For an overview of the new features and enhancements in TR7, see the announcement letter at IBM i 7.1 Technology Refresh 7 offers performance, usability, and integration enhancements. More detailed technical descriptions can be found at IBM i Technology Updates.


If you are running SAP on IBM i, the following enhancements may be of special interest for you:


  1. IBM i 7.1 Technology Refresh 7 offers support for several new hardware features, such as solid-state drives (SSDs). For a complete list, see the anouncement letter that we link to above.
  2. The maximum size of an SQL index has been increased to 1.7TB. Non-partitioned tables could already grow up to 1.7TB, but for indexes, the size was still limited to 1 TB. This limit has now been increased.
  3. Phase 2 of the tracking of important system limits has been shipped. In our blog entry of October 25th 2013, we had already announced that: SAP on IBM i - Update week 43 2013: Proactively monitoring table size limits on IBM i. The following article provides a very good and detailed overview over the concept: OnDemand Tracking of Important System Limits on IBM i. In the meantime we found out, that you will only see an entry with SIZING_NAME = 'MAXIMUM NUMBER OF VARIABLE LENGTH SEGMENTS'  in view QSYS2/SYSLIMITS if you have a table that actually has more than 100 variable length segments. So if you don't see such an entry after applying the technology refresh, do not worry: You just don't have a table that is coming close to the critical limit.


The database enhancements in IBM i 7.1 technology refresh 7 are mostly available through DB2 PTF group SF99701 level 26 as well. Those enhancements will also be made available for IBM i 6.1 through DB2 PTF group SF99601 level 31, which is scheduled to become available on December 6th 2013. However, other features of IBM i 7.1 TR7, in particular the support of certain hardware components, may not become available for IBM i 6.1 at all.

Welcome to the SAP on IBM i update! I'd like to draw your attention to an upcoming event.


Meeting of the “SAP on IBM i” workgroups of the German SAP User Group (“Deutschsprachige SAP Anwendergruppe e.V. / DSAG”) and the German IBM User Group (“COMMON Deutschland e.V.”)


The DSAG is an SAP customer forum for German speaking customers in Germany, Austria and Switzerland and it is the largest SAP user group worldwide. For details please refer to https://www.dsag.de/dsaghome/dsag.html.
The COMMON Deutschland  is an IBM customer forum for Germany and is member of COMMON Europe. For details please refer to http://www.common-d.de.

On November 14th 2013, the “SAP on IBM i” workgroups of the DSAG and COMMON Deutschland get together for a one-day meeting packed with interesting presentations and discussions around SAP on IBM i. For details on the agenda and for registration purposes please refer to https://www.dsag.de/dsaghome/veranstaltungen/uebersicht/details/veranstaltung/dsagveranstaltung/20131114/list-88/kalender/AG_SAP_on_IBM_i_Treffen-4.html or http://www.common-d.de/index.php/component/eventlist/details/14-arbeitskreis-sap-on-ibm-i?Itemid=435 respectively.


Have a good time and see you later!

Like any other database, DB2 on IBM i has certain limits for maximum object size, maximum number of rows and many more attributes of database objects, such as tables and indexes. The limits are not always the same, some of the limits are being raised with newer releases of IBM i. The current SQL and database limits for IBM i 7.1 are documented here: SQL limits.


As you can see in the infocenter documentation, the maximum size of a non-partitioned table and the maximum size of a data partition are shown as 1.7 terabytes. However, we recently discovered another limit that should be proactively monitored on tables containing columns with variable length data, such as BLOB, CLOB, DBCLOB, VARCHAR and VARGRAPHIC columns. When the variable length data makes up a significant share of the overall table data, the maximum number of variable-length segments may be exhausted before the size limit of 1.7 terabytes has been reached.  It should be very rare that one encounters this limit, however it can be checked to ensure that this limit, and hence an error SQL0904 type 6, is avoided.


How do I know if my largest tables are affected?


On October 7, 2013 IBM has announced IBM i 7.1 technology refresh 7 (see IBM i 7.1 Technology Refresh 7 offers performance, usability, and integration enhancements). General availability for IBM i 7.1. TR7 is planned for November 15, 2013. With this technology refresh, the tracking of important system limits (catalog view QSYS2/SYSLIMITS) will be enhanced to show when tables reach the maximum number of variable-length segments. You can see more details here: Tracking important system limits. This function will also be made available for IBM i 6.1 at a later time.


We recommend installing IBM i 7.1 TR7 to be able to check these limits in an easy way.


If you wish to check these limits prior to installing IBM i 7.1 TR7, you can still find out if you have any tables which may be close to the limit by performing the following steps on tables that are close to or above 1 terabyte in size by following the manual steps below.


  1. Identify large tables in your SAP application server ABAP by using transaction DB02 or DBACockpit -> Space -> State on Disk. Select Object Details, leave the wildcard character '*' for the Table Selection and select the Update function. To see the largest tables on top of the list, select the column labeled as Size (KB) and sort the data in descending order.
  2. To map the SQL name(s) of the largest table(s) to the object name(s) at the operating system level, you can use interactive SQL (STRSQL or the System i Navigator) and execute the following query:
    WHERE TABLE_NAME = '<SQL name>'
  3. With the object name, enter the following command at a command line:
    This will create a spooled file named QPSRVDMP in your job.
  4. Use command WRKSPLF and option 5 to display the spooled file QPSRVDMP. Search for a string *QDDS using function key F16 repeatedly until you see the following text:
    OBJECT TYPE-           DATA SPACE                                      *QDDS
    NAME-        <object>  <object>              TYPE-          0B   SUBTYPE-          90
    CREATION-    09/30/05  15:03:57              SIZE-          0052060000
    OWNER-       <SID>OWNER                      TYPE-          08   SUBTYPE-          01
    ATTRIBUTES-          0800                    ADDRESS-       146D46E631 000000
    Make sure that you see the 0B as type and 90 as subtype - the search will also find other places where the object type is *QDDSI  instead of *QDDS. Note the value to the right of the text ADDRESS, or copy it into the clipboard (in the example: 146D46E631 000000).
  5. Use the command STRSST to start the system service tools and sign on with a service tools user. Select the following menu items in the given order:
    1. Start a service tool
    4. Display/Alter/Dump
    2. Dump to printer
    1. Machine Interface (MI) object
    11. Data space (0B)
    2. Find by object address
  6. Type or copy the address that was found in the previous step (in the example: 146D46E631 000000) and hit Enter. On the next screen ("MI Object Found") press Enter again.
  7. On the screen "Select Format", select option 2. Base structure. Specify a dump title (free text) and hit Enter. After that, you will see a message "Dump to printer successfully submitted" at the bottom of the screen. The created spooled file has the name QPCSMPRT.
  8. Leave the system service tools by using function key F3 multiple times.
  9. Use command WRKSPLF and option 5 to display the spooled file QPCSMPRT. Search for a string #AUXS using function key F16. You will see a text similar to the following:
    LAST1            000000000000AAAA        AUXT@            146D46E631 000DF
    #AUXS            00000040                #AXSL            00000057
    #AXEN            00000000000C0785        AXLUB            0B8DF2796E 70F60
    The hexadecimal number to the right of the text #AUXS indicated the number of variable-length segments currently in use, in the example 00000040. If this value approaches 0000FFFD, you are likely to hit the described table size limit.


What do I do if I approach or reach the limit?


If the critical table contains a lot of deleted rows, you may be able to free up space by reorganizing the table as described in SAP Note 84081. Please note that the space occupied by variable-length data can only be freed completely during an offline reorganization. While the offline reorganization is running, you may temporarily need twice the space of the table available on disk.


If reorganizing the table does not help, you can either archive some of the data and delete it from the table, or you can use table partitioning in order to distribute the data over multiple partitions that can each have the limit of one non-partitioned table. Background information about table partitioning in BW systems can be found in SAP Note 815186. Partitioning of non-BW tables should only be implemented after consulting the SAP on IBM i porting team. If you are planning to do that, please open a ticket for component BC-DB-DB4.


The above information will also be made available through SAP Note 1930962.

Hello fellow SAP on IBM i people,


I'd like to point out a newly published technical reference of a large scale SAP HA on IBM i implementation at Al Mansour Holding Company.



SAP on IBM i installations are often designed with Disaster Recovery (DR) in mind. Many DR configurations are based on having the database and file system in an independent mirrored storage pool (iASP) or mirrored by logical replication. By configuring an SAP system to also use replicated central services according to the SAP HA standard, the SAP system can be switched between hosts in near zero downtime for maintenance or in the case of failure.


The reference describes the technical details of how Al Mansour Holding Company operates three Highly Available SAP clusters for ERP, SCM, and BW based upon application HA with replicated SAP central services, database and file system mirroring with iASP, and managment by PowerHA for IBM i.


Click the following link to be directed to the IBM site where the reference is published: http://www.ibm.com/software/success/cssdb.nsf/CS/STRD-9BTH2H; alternatively, click here to be directed to the PDF itself: https://ibm.biz/BdDzQL


By the way, if you are also considering installing or converting an SAP system to an Highly Available one based on SAP central services, or have further questions, we'd like to know; please open a ticket to BC-OP-AS4 and tell us.



    ... Eric

Welcome at the SAP on IBM i update! This time we will focus on the replacement of the ASCII NW RFC SDK in ILE by a UNICODE version:


Many of you already have recognized that there are several versions of the NW RFC SDK available for OS/400, while there is normally only one version for all of the other platforms. This standard version is a UNICODE one which we also ship (NWRFCU_xxx.SAR). Because most of you are not familiar in writing PASE c-Programms to link against this standard version we also ship some ILE versions of the NW RFC SDK.

The easiest version to use is the EBCDIC one (NWRFC_E-xxx.SVF). It allows to simply write a native ILE PGM to link against this version. But this version has many restrictions. It only accepts the latin characterset specified by the SAP Codepage 0120. For this reason we also supported an ASCII Version (NWRFC_A-xxx.SVF) as we are used to from the classic RFC SDK. Unfortunately this does not really work for the NW RFC SDK. So we were urged to ship an additional UNICODE Version (NWRFC_U-xxx.SVF).

This version has to be used in any case where the SAP Codepage 0120 is not sufficient. The ASCII version is now obsolet and will never be patched again.

Unfortunately the UNICODE version requires a full UNICODE handling of the SDK library interfaces. Every imported/exported string to the library must and will be UNICODE and the RFC Programm has to handle this.


For more information, please see SAP Note 1097997 'Installing SAP NW RFC SDK on IBM i' .


Stay tuned...

As the development team intends to maintain a consistent format, editions of the series will continue to be published. Set feeds accordingly if you are interested in following the series

Over the years this IBM i event at SAP TechED has become a tradition.


Save the date  (Monday October the 21st.)  and register now free of charge and reserve your seat for our annual IBM i and SAP Annual Solution Update and Dinner at TechEd 2013 Las Vegas.


IBM and SAP experts will provide the latest news and solution updates for IBM i and SAP environments. Attend this session and learn about IBM PureFlex Systems, SAP solution updates, and exciting new possibilities with cloud computing and In-Memory Computing. Don't miss this opportunity to have your questions answered by experts and to engage in lively discussions with your peers.


The business meeting will be followed by a group dinner.


For further details please follow the link



In the past, the installation master media was always bundled together to the corresponding product media. With the recent introduction
of the Software Logistic Toolset (SL Toolset) 1.0, customers might wonder:
What tool do I need to use when executing installation related activities?
The SL Toolset provides tools for distinct software logistic areas: Software Maintenance (formerly Upgrades/Updates), System Provisioning (formerly
Installations), Change Controls and Software Automation. This blog entry will help you understand how to find and download the correct tool and version for
the following system provisioning tasks:
  • Installation & system copy for products based on SAP NetWeaver – Software Provision Manager (SWPM*.SAR / 70SWPM*.SAR)
  • Split of Dual-Stack Systems into separate ABAP and Java Systems - Dual-Stack Split Tool (DSS21SP00*.SAR / 70DSS21SP00*.SAR)
  • Adapt key characteristics of an SAP system - System Rename (SYSRENAME*.SAR / 70SYSRENAME*.SAR)
These system provision tools are all based on the well-known SAPinst framework. The Dual-Stack Split and System Rename tools will become part of the SWPM tool starting with SWPM 1.0 Support Package Stack 09 in Q4 2013 and thus will no longer be available for download. They are the only currently supported tools for SAP software logistic installation tasks and delivered as part of the Software Logistic Toolset available for download at the SAP Service Marketplace (http://service.sap.com/sltoolset).
How to find out thecorrect tool version?
All toolkits are available in two (2) different versions: aversion is for systems based on SAP NetWeaver 7.0x and a version for systems based on SAP NetWeaver 7.1 and higher.  Therefore, to find out which tool is needed for your specific SAP system, it is critical to know the SAP NetWeaver technology release for the (to be) installed system.
The releases and it tool name are:
ReleaseTool NameSupported Products
SAP systems based on NetWeaver 7.1 and higher including:
  • SAP Solution Manager 7.1 on HANA
SAP systems based on NetWeaver 7.0x including:
  • SAP Solution Manager 7.1 SR 1
  • SAP Solution Manager 7.0 including EhP 1 SR 1
A full list of products and its corresponding toolkit version is available in SAP Note 1680045. The SAP NetWeaver release for a product (i.e. EHP6 FOR SAP ERP 6.0) can be found in the installation guide or download center for example. For more information, see the SAP Product Availability Matrix
Additional details:
  • All new installations, system copy imports, rename and dual-stack split operation require the usage of SAP kernel release 720(_EXT) or higher. A link to the 720 kernel to be used with SWPM installation can be found on the SL Toolset webpage.
  • We recommended that you use the latest SL Toolset Stack when you start your project. The maintenance strategy is clearly explained in the SL Toolset webpage and the corresponding SAP release note.
  • SL Tools guides are constantly updated and also available on the SL Toolset webpage.

With this update, we would like to inform you about a reference story that has recently been published by IBM.


As you might know, IBM PureFlex Systems offer a tight integration of compute, storage and networking resources to simplify deployment and management while reducing power, cooling and operational cost. PureFlex Systems are pre-configured, pre-integrated solution with x86 or POWER processors.


Recently IBM published a reference story called "Averna S.p.A. upgrades their SAP IT environment to increase compute and networking performance while reducing costs". Averna installed SAP on IBM i and consolidated their Windows workloads onto the IBM PureFlex System and increased their storage performance and capacity with Storwize V7000. The SAP application performance has improved by up to 200 % while IT power consumption has been cut by 30%.


Read the full reference story here: http://public.dhe.ibm.com/common/ssi/ecm/en/spc03456wwen/SPC03456WWEN.PDF

Welcome to the SAP on IBM i update! You will have recognized that our team started a new effort to maintain this blog. By distributing the actual work over more members of the team we hope to reach better continuity.


This time I'd like to return to kernel maintenance. I know we talked about this several times already. But there is one detail, which I feel every admin needs to take into his or her considerations.


How it used to work...

What was a proper way to patch a kernel in the old days, when everything was built up from a kernel library? I suppose most of you did something similar to creating a copy of the library and applying a patch which APYR3FIX to this library. The library was a container that could be transported to any other system for or after testing. The actual update of the target system required downtime to run APYSAP (or its predecessor APYR3KRN) in order to create or update the corresponding file system objects. So preparation could be taken offline.


... and how it can work today

At latest with activating the downward compatible kernel 7.20 you faced  a paradigm shift of how the objects in file system and in the kernel libraries are deployed. Deployment has to take place into the master directory DIR_CT_RUN. All convenience and safety actions of tool APYSIDKRN take place while the patch packages are examined. This includes saving off the existing kernel objects, checking for the right sequence of the packages that are deployed and many more and takes time - you probably do not want all this to happen while your system is down.


And here is my point: If your system is not configured for replication with sapcpe your instances are running on DIR_CT_RUN. All the described actions including potential errors in the process that need a fix will have to run during downtime. With replication, where every instance of an SAP system has its own copy of the kernel, this copy is updated automatically and quick at system start, while the expensive stuff can run offline for the master directory.


See Note 1632754 - IBM i: Changing to an instance-specific directory to get into the details.

Addendum July 30, 2013: Document SAP on IBM i Kernel Maintenance deals with the same topic

Welcome to the SAP on IBM i update! With this update, we would like to introduce you to profile parameter ES/TABLE, which can be used to switch the shared memory implementation on IBM i.


While ES/TABLE=UNIX_STD is the default and used to be significantly faster in IBM i releases prior to IBM i 6.1, the alternative ES/TABLE=SHM_SEGS can be significantly faster in IBM i releases 6.1 and higher when hardware with a large number of processors is used. We have seen significant improvements on systems with more than 16 processors.


Before you can change your SAP system to ES/TABLE=SHM_SEGS, you must ensure that your system is at IBM i 6.1 or higher, install the downward-compatible kernel 7.20 patch at level 220 or above and switch to the new user concept (see SAP Note 1149318). When changing the profile parameter ES/TABLE to SHM_SEGS, you also need to consider profile parameters ES/SHM_BASE_ADDR and as4/MAXSHR64. Profile parameter ES/SHM_BASE_ADDR defines the starting address for extended memory, and its value should be large enough to avoid overlapping with the SAP buffers. Profile parameter as4/MAXSHR64 defines the maximum number of 256 MB segments, and its value must be large enough so that the resulting addresses calculated on top of ES/SHM_BASE_ADDR fit into the address range that is spanned by as4/MAXSHR64. For a better understanding of the calculations, you might want to look at the presentation at https://scn.sap.com/docs/DOC-44113.


To calculate a good value for ES/SHM_BASE_ADDR, you first need to find the highest shared memory address used by your SAP buffers. You can find that by executing ABAP program RSTUNSHM in transaction SE38 and sorting the output list by the address. Depending on your buffer settings, you will get an address such as 07000002A00829E0. You must generously round up to a free address, in this case we recommend ES/SHM_BASE_ADDR=0x0700001000000000 (the prefix '0x' indicates a hexadecimal number).


You now need to consider the value for as4/MAXSHR64. The default value of 256 represents 64 GB (256 * 256 MB). If you calculate the highest possible address for 64 GB, you end up with a hexadecimal value of 0x0000001000000000. If you map this to the value of ES/SHM_BASE_ADDR, you can see that the maximum allowed distance from the segment start (defined by as4/MAXSHR64) is the same as the beginning of the extended memory (defined by
ES/SHM_BASE_ADDR). If you try to start your system with these settings, you will receive errors very soon. In order to allow more extended memory, you need to set as4/MAXSHR64 to a value higher than 256, for example as4/MAXSHR64=2048. This value represents 512 GB (2048 * 256 MB), and with that the highest possible distance from the segment start is 0x0000008000000000, so you have enough address space for up to 448 GB of extended memory. Note that the space is not actually allocated on the system unless it is really needed. These values only represent addresses, so setting as4/MAXSHR64 too high usually does not hurt (it was intended as a sort of protection in case of kernel errors).


Profile parameter ES/TABLE and its related parameters are also discussed in SAP Note 808607 in the section 1.1.8. “Improvements as of IBM Version i6.1”.

Welcome to the SAP on IBM i update! With this update, we would like to inform you about a whitepaper that has recently been published by IBM.


As you know, the storage footprint of an SAP system tends to grow over time. This will not only impact overall system performance, but also slow down maintenance tasks like backups and SAP release upgrades. There are many options available to reduce the storage requirements of an SAP system. Some options like housekeeping are provided by the software (SAP and IBM), others – like disk compression – are provided by the hardware.


The white paper "Exploit Storage Resources for SAP on IBM i" summarizes the available hardware and software options, briefly explains how they impact storage resources and provides references to detailed instructions on how to use them. You can find the whitepaper at http://www.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102202.


Filter Blog

By author: By date:
By tag: