Service Report SAP EarlyWatch Alert SAP System ID ABC SAP Component NW Release 04 Status Productive DB System ORACLE 9.2.0.8.0 Customer Service Center Telephone Fax Date of Session Date of Report Author 8.02.2007 8.03.2007 EarlyWatch Alert Service <Watermark> <dat_dummy> Session No. Installation No. Customer No. Service Summary 1 Service Summary During this EarlyWatch Alert Session, we detected potential problems concerning your system. We recommend that you take corrective action as soon as possible. Note: All recommendations provided in this report are based on our general experience only. We advise you to test our recommendations before using them in your production system. Also note that EarlyWatch Alert is an automatic service. Section Overview Rating Checked Area System Configuration System Performance Workload Distribution SAP System Operating Database Settings Database Administration BW Checks Security Checks Note: For more information about EarlyWatch Alert, a sample EarlyWatch Alert report with explanations is available in the Media Library on SAP Service Marketplace (http://service.sap.com/ewa). This provides an overview of the check rating strategy and the KPIs that trigger the EWA alerts. Alert Message Overview Priority Description New Alert Medium There are security issues in your system. New Medium Your database is growing very rapidly. New Note: If you need help investigating the alerts listed, order an EarlyWatch contract by contacting your local support organization, or by creating a customer message in component XX-SER-TCC. If you already have such a contract, a detailed analysis can be performed during the next Service Session. Please address the topic during the preparation of the session. Note: If you experience serious performance problems, create a customer message on component XX-SERTCC with priority 'high' or 'very high'. EarlyWatch Alert Report , , 8.03.2007 2 Performance Indicators 2 Performance Indicators The following table shows the relevant performance indicators in various system areas. Area System Performance Hardware Capacity Database Space Management Query Performance Indicators Active Users Max. CPU Utilization on DB Server Max. CPU Utilization on Appl. Server DB Size Last Month DB Growth Avg. Total Runtime of the BW Queries Avg. Database Runtime of the BW Queries Value 18 74 % 74 % 355.52 GB 118.63- GB 11.5 s 8.0 s Trend down steady steady steady steady down steady 3 Recommended Service The analysis of your SAP EarlyWatch Alert service session has shown that your installation has a high optimization potential. Therefore, SAP would highly recommend you to perform following service: Service Issue Scope Benefit SAP Security Optimization Service Special BI Service Some users have critical authorizations in certain transactions A special BI service based on the BI Key Performance Indicators is recommended Increase of security in your SAP System Avoids severe security problems in your SAP solution BI Service type dependent To find details about this BI service, please have a look at chapter 'BI - KPIs' Average Consulting Days 1 4 System Configuration To ensure optimal support of your SAP system, the SAP components, database, and operating system used must all be supported. Detailed information about SAP's platform support is available on SAP Service Marketplace at http://service.sap.com/platforms. Rating Check Hardware Configuration Software Configuration SAP Service Preparation 4.1 Hardware Configuration The table below shows the detected hardware configuration: HARDWARE CONFIGURATION Server Type Server Others usnasap101 Database Server usnasap123 DATABASE SYSTEM Database Server usnasap123 OS AIX AIX OS Version 5.3 5.3 Database System ORACLE Number of CPUs 8 14 Memory [MB] 8192 18432 Current Version 09.2.0 4.2 Software Configuration The table below shows the installed software components. The "Rating" column is filled only for the SAP R/3 PlugIn (PI) and SAP Basis Plug-In (PI_BASIS), and is based on the maintenance schedule. COMPONENTS Component BI_CONT PI_BASIS SAP_ABA Component Release 353 2005_1_640 640 EarlyWatch Alert Report , , 8.03.2007 Patch Level 0003 0003 0013 Rating 3 System Configuration Component SAP_BASIS SAP_BW ST-A/PI ST-PI Component Release 640 350 01H_BCO640 2005_1_640 Patch Level 0013 0013 0000 0004 Rating 4.2.1 SAP Kernel Release Your SAP kernel release (release 640, patch level 151) is not up to date. Recommendation: Ensure that you are using the latest SAP kernel. See SAP Notes 19466 and 138704. (These notes refer to further SAP Notes, depending on your system characteristics.) 4.3 SAP Service Preparation For detailed information about the required steps, see SAP Note 91488 (SAP Support Services: Central preparation note). Please ensure that all the functionality mentioned in the SAP Note is up to date, especially connections, collectors, and service tools. 4.3.1 SAP Service Tools Before we can deliver any services, the latest version of the SAP Service tools must be available in your system. The last run of report RTCCTOOL was 20.11.2006. Recommendation: You should run report RTCCTOOL, as described in the SAP Note 91488, to update tool release information. In case of missing transports detected by RTCCTOOL, implement the recommendations given by the tool. EarlyWatch Alert Report , , 8.03.2007 4 System Configuration 4.4 BW Workload 4.4.1 Workload per User and Navigation Steps This overview takes into account the following: - The number of users who execute queries (grand total) - This number is grouped according to InfoConsumer, Executive, and Power User (totals), depending on their number of navigation steps - The InfoConsumer is divided again according to the number of navigation steps (subtotals). - The timeframe is the last complete week from Monday to Sunday. User/Consumer Grand total: Users performing queries Total: Info Consumer [1 - 400 Nav Steps/ week] ...Sub total: Info Consumer 1-10 Nav Steps/ week ...Sub total: Info Consumer 11-50 Nav Steps/ week ...Sub total: Info Consumer 51-100 Nav Steps/ week ...Sub total: Info Consumer 101-200 Nav Steps/ week ...Sub total: Info Consumer 201-300 Nav Steps/ week ...Sub total: Info Consumer 301-400 Nav Steps/ week Total: Executive [401 - 1200 Nav Steps/ week] Total: Power User [> 1200 Nav Steps/ week] Number 92 91 29 41 10 8 2 1 0 1 4.4.2 Reporting and Upload Workload last week The diagram above shows an overview of the workload distribution regarding reporting and upload activities from the last week. Note that the values shown do not reflect the actual values. In each case, we have taken the highest value and considered it to be "100". The other values show the ratio to the maximum values. Maximum values are given below. Note that the minimum requirement is ST-A/PI 01G*. If this has not been applied, no reporting activity will be shown in the diagram. max. value # Nav. Steps 422 max. value # Uploads 2 4.5 History of Activity and Average Response Times The following graphics or tables show the system activity, which is the number of dialog steps and the average response times for the past 20 weeks. EarlyWatch Alert Report , , 8.03.2007 5 BW Runtime Statistics The top table or graphic compares the total system activity to the dialog activity in the time window specified. The bottom table shows the average CPU, DB, and wait times as parts of the average response time for dialog tasks in the time window specified. 5 BW Runtime Statistics The performance of your queries and upload was analyzed with respect to average runtime and total workload. The following table provides an overview of your system activity and performance from the BW point of view. Note: All queries using the 'Read API' of your system (such as from connected SAP-APO or SAP-SEM systems) are named 'RSDRI_QUERY,' so you cannot locate them in your BW system. Please note that the following chapters only contain queries/InfoCubes for which the statistics indicators are set. 5.1 Top InfoProvider The following table lists the top 5 InfoProviders regarding the number of uploaded records. TOP INFOPROVIDERS PER UPLOAD VOLUME InfoProvider # InfoSources 0BWTC_C02 1 # Requests 1 # Records written 452 Total Time [sec] 37.6 The following table lists the top five InfoProviders based on the number of query hits. EarlyWatch Alert Report , , 8.03.2007 6 BW Runtime Statistics TOP INFOPROVIDERS PER NUMBER OF QUERIES InfoProvider Navigation Avg. Steps runtime [s] YMPAPL_1 YM13_INV YM_PTFINV ZMORDINV YMPAPO_1 1108 1420 2756 1032 5 21.60 16.20 7.00 3.90 418.60 Runtime [%] Avg. time OLAP [s] 33 32 27 6 3 10.70 0.50 0.10 0.10 1.70 Avg. time Avg. DB [s] Frontend time [s] 9.50 0.90 15.00 0.60 5.60 0.30 1.30 0.30 412.80 0.90 Avg. # Frontend cells 0 1 448 1147 0 5.2 Frontend Distribution # Queries 47 BEX 1 WEB 42 ODBO 4 API 0 The table above provides an overview of the front-end distribution. It contains the total number of queries executed over the last complete week (Monday to Sunday) and the number of queries executed from the different front ends. 5.3 Query Profile Check 5.3.1 Queries The table below gives an overview of the query runtimes in regard to the use of the OLAP Cache. This means the line 'Cache Queries' summarizes all the queries which used the OLAP Cache, whereas 'DB Queries' directly accessed the database. Please note that all queries using the 'Read API' (Queries named 'RSDRI_QUERY') cannot use the OLAP Cache. Therefore these queries are listed separately. If there were no queries in the last seven days which used the OLAP Cache, accessed the database or used the 'Read API', the corresponding summary line is not displayed. Task Type All Queries DB Queries Cache Queries Nav. Steps 6676 4077 2599 Runtime > 20 seconds [%] 5 8 1 Avg. Runtime [s] 11.54 16.27 4.11 Avg. Time OLAPINIT [s] 0.24 0.21 0.30 Avg. Time OLAP [s] 2.02 1.31 3.13 Avg. Time DB [s] 8.01 13.12 0.00 The following table provides a summary of query runtimes and distinguishes between the different front ends. (BEX, WEB, ODBO and API). If no queries were started with one of the given options over the last seven days, the corresponding summary line is not displayed. Task Type All Queries WEB Queries Nav. Steps 6676 3727 Runtime > 20 seconds [%] 5 9 EarlyWatch Alert Report , , 8.03.2007 Avg. Runtime [s] 11.54 17.45 Avg. Time OLAPINIT [s] 0.24 0.21 Avg. Time OLAP [s] 2.02 3.57 Avg. Time DB [s] 8.01 12.90 7 BW Runtime Statistics Task Type BEX Queries ODBO Queries Nav. Steps 3 Runtime > 20 seconds [%] 0 Avg. Runtime [s] 6.30 Avg. Time OLAPINIT [s] 0.27 Avg. Time OLAP [s] 0.30 Avg. Time DB [s] 1.70 2946 1 4.06 0.29 0.07 1.84 5.3.2 Top Time Queries by Total Workload The total workload caused by queries is defined as the sum of the total runtimes of all queries. The following query profile lists the queries, as a percentage of total runtime, that contribute the greatest amount to the total workload. Query name InfoCube Total ZYMPAPL_1_Q0001 ZYM13_INV_Q0007 ZYM_PTFINV_Q0001 ZYM_PTFINV_Q0001_FS ZYMPAPL_1_Q0006 ZYM13_INV_Q0008 ZZMORDINV_Q0001_FSF ZYM13_INV_Q0015 ZYMZO_1_Q0008 YYMPAPO_1_Q0002 YMPAPL_1 YM13_INV YM_PTFINV YM_PTFINV YMPAPL_1 YM13_INV ZMORDINV YM13_INV YMZO_1 YMPAPO_1 Nav. Runtime Avg. Steps [%] runtime [s] 5692 875 546 669 2040 124 391 846 143 56 2 100 27 21 17 10 7 6 5 3 2 2 12.12 21.29 26.63 17.53 3.51 40.33 10.08 3.96 14.83 25.76 555.18 Avg. DB time [s] 8.29 11.74 25.30 16.69 1.96 0.66 8.57 1.03 13.83 10.07 543.10 Avg. OLAP time [s] 2.23 8.00 0.35 0.22 0.05 38.44 0.85 0.04 0.49 1.04 3.97 Avg. Frontend time [s] Avg. # Frontend cells 0.54 0.87 0.86 0.52 0.28 1.18 0.44 0.27 0.39 5.40 1.06 1786 0 0 0 606 0 0 1399 0 138307 0 5.3.3 Top Time Queries by DB Load The total database workload generated by the BW system is the sum of the total database access times of all queries. The following query profile lists the queries, as percentages of total database access time, that make up the largest part of the database load. Query name Total ZYM13_INV_Q0007 ZYM_PTFINV_Q0001 ZYMPAPL_1_Q0001 ZYM_PTFINV_Q0001_FS ZYM13_INV_Q0008 ZYM13_INV_Q0015 YYMPAPO_1_Q0002 YYMPAPO_1_Q0001 ZYM13_INV_Q0017 ZZMORDINV_Q0001_FSF InfoCube Navigations Db load in [%] Avg. DB time [s] Avg. Runtime [s] 5519 100 8.77 11.68 YM13_INV 546 29 25.30 26.63 YM_PTFINV 669 23 16.69 17.53 YMPAPL_1 875 21 11.74 21.29 YM_PTFINV 2040 8 1.96 3.51 YM13_INV 391 7 8.57 10.08 YM13_INV 143 4 13.83 14.83 YMPAPO_1 2 2 543.10 555.18 YMPAPO_1 3 2 325.99 327.59 YM13_INV 4 2 221.59 223.60 ZMORDINV 846 2 1.03 3.96 5.3.4 Top Time Queries by Average Runtime The ten queries whose average runtimes have the highest optimization potential are listed here. Query name InfoCube Avg. Runtime [s] Total 34.43 YYMPAPO_1_Q0002 YMPAPO_1 555.17 ZYO13_ITM_Q0003 YO13_ITM 375.17 YYMPAPO_1_Q0001 YMPAPO_1 327.59 ZYM13_INV_Q0017 YM13_INV 223.60 ZYM_OO_Q0002 YM_OO 165.34 ZZM_MAT_Q0002 ZM_MAT 149.65 ZYMPAPL_1_Q0006 YMPAPL_1 40.33 ZZMORDINV_Q0002 ZMORDINV 28.09 ZYM13_INV_Q0007 YM13_INV 26.63 ZYMZO_1_Q0008 YMZO_1 25.76 EarlyWatch Alert Report , , 8.03.2007 Avg. DB time [s] 24.75 543.10 372.95 325.99 221.59 65.34 115.61 0.66 27.37 25.30 10.07 Avg. OLAP time [s] 7.39 3.97 0.61 0.17 0.66 77.94 10.26 38.44 0.29 0.35 1.04 Avg. Frontend Avg. # time [s] Frontend cells 1.24 10368 1.06 0 0.44 0 0.73 0 0.43 0 0.74 0 0.32 0 1.18 0 0.39 0 0.86 0 5.40 138307 8 BW Runtime Statistics 5.4 BEX - Queries by Total Workload (TOP 10 per Runtime%) The table below contains data about the 10 queries that consumed the most time, with regard to runtime. Note that this table contains data about single executions of your queries. This means the data is not summarized, and the name of a query may appear several times. Query name InfoCube Nav. Steps Runtime [%] Avg. runtime [s] 3 3 100 100 6.30 6.30 Total ZYM13_INV_Q0040 YM13_INV Avg. DB time [s] 1.70 1.70 Avg. OLAP time [s] Avg. Frontend time [s] Avg. # Frontend cells 0.30 0.30 3.94 3.94 455 455 5.5 WEB - Queries 5.5.1 WEB Templates (TOP 10 per Runtime) The following table lists the top 10 templates that contribute the greatest amount to the total workload. Template Name (Page ID) Runtime # # Avg. Max. Avg. Avg. [s] Executions Navigation runtime runtime Template Template [Sessions] Steps for one for one prepare render [Queries] NavStep NavStep time [s] time [s] [s] [s] Total 65019.10 3725 3727 17.45 2051.11 0.32 0.34 <undefined> 64384.99 3395 3397 18.95 2051.11 0.27 0.31 ZEXCELPRINT_PORTRAIT 324.88 158 158 2.06 20.95 0.79 0.59 ZEXCELPRINT_LANDSCAPE 264.09 148 148 1.78 6.19 0.66 0.69 ZMIG_BW_PRINTFITTO_TMPL 45.14 24 24 1.88 5.16 1.23 0.10 5.5.2 WEB - Queries by Total Workload (TOP 10 per Runtime%) The table below contains data about the 10 queries that consumed the most time, with regard to runtime. Note that this table contains data about single executions of your queries. This means the data is not summarized, and the name of a query may appear several times. Query name InfoCube Total ZYMPAPL_1_Q0001 ZYM13_INV_Q0007 ZYM_PTFINV_Q0001 ZYMPAPL_1_Q0006 ZYM13_INV_Q0008 ZYM13_INV_Q0015 YYMPAPO_1_Q0002 YYMPAPO_1_Q0001 ZYM_OO_Q0002 ZYM13_INV_Q0017 YMPAPL_1 YM13_INV YM_PTFINV YMPAPL_1 YM13_INV YM13_INV YMPAPO_1 YMPAPO_1 YM_OO YM13_INV Nav. Runtime Steps [%] 2759 875 546 669 124 391 143 2 3 2 4 100 31 24 20 8 7 4 2 2 2 1 Avg. Avg. Avg. Avg. runtime DB OLAP Template [s] time [s] time [s] prepare [s] 21.71 15.95 4.71 0.33 21.29 11.74 8.00 0.39 26.63 25.30 0.35 0.35 17.53 16.69 0.22 0.32 40.33 0.66 38.44 0.43 10.08 8.57 0.85 0.21 14.83 13.83 0.49 0.15 555.18 543.10 3.97 0.62 327.59 325.99 0.17 0.58 482.78 196.02 233.81 0.54 223.60 221.59 0.66 0.30 Avg. Template render [s] 0.38 0.48 0.50 0.19 0.75 0.23 0.24 0.44 0.15 0.21 0.13 5.5.3 WEB Templates - Multiple Query/View usage Several Web items can use the same data provider if there is no need for independent navigation in a template. Based on our experience, independent navigation is not required in most cases, and you can therefore avoid the multiple query execution by assigning a query or view to just one data provider, which is used by all web items. 5.5.4 WEB Templates - Design Statistics TEMPLATE COMPLEXITY REGARDING DIFFERENT QUERIES/VIEWS # of Queries / Views 01 02 03 04 05 06 07 08 EarlyWatch Alert Report , , 8.03.2007 # of Web Templates 0 0 0 0 0 0 0 0 9 BW Runtime Statistics # of Queries / Views 09 10 > 10 # of Web Templates 0 0 0 5.5.5 WEB Templates - Unused Data Providers TOP10 NUMBER OF UNUSED DATA PROVIDERS PER TEMPLATE Template ZMIG_BW_CENTERFRAME # unused Data Providers 1 TOTAL NUMBER OF UNUSED DATA PROVIDERS Total number of unused Data Providers 1 Recommendation: Unused data providers consume OLAP initialization runtime and can cause transport problems. In particular, if you have several templates that contain unused data providers, we recommend that you delete them to obtain an optimized and clean template implementation. Using BW 3.x, unused data providers must be deleted manually from the HTML code in the Application Designer. As of the BI in SAP NetWeaver 2004, there is a delete function in the data provider's context menu. 5.6 ODBO - Queries by Total Workload (TOP 10 per Runtime%) The table below contains data about the 10 queries that consumed the most time, with regard to runtime. Note that this table contains data about single executions of your queries. This means the data is not summarized, and the name of a query may appear several times. Query name InfoCube Total ZYM_PTFINV_Q0001_FS ZZMORDINV_Q0001_FSF ZYMZO_1_Q0008 ZYM_OO_Q0002 YM_PTFINV ZMORDINV YMZO_1 YM_OO Navigation Runtime Avg. steps [%] runtime [s] 2946 2040 846 56 4 100 60 28 12 0 4.06 3.51 3.96 25.76 6.62 Avg. DB time [s] 1.84 1.96 1.03 10.07 0.00 Avg. OLAP time [s] 0.07 0.05 0.04 1.04 0.00 Avg. ODBO Time [s] 0.10 0.10 0.08 0.38 0.39 Avg. Prepare Axis [s] 0.15 0.02 0.07 5.94 5.26 5.7 Analysis of MultiProvider Queries During parallel processing of MultiProvider queries, subprocesses that process the partial accesses to the InfoProviders in the MultiProvider are separated from a parent process. The parent process provides a synchronization point where the overall result is collected. In contrast to sequential processing, which permits partial results to reach the OLAP processor, parallel processing requires that the overall result first be collected at the synchronization point. To avoid an overflow of the process memory, parallel processing is cancelled as soon as the collected result contains 30,000 rows or more and there is at least one subprocess not finished. The MultiProvider query is then restarted automatically and processed sequentially. This strategy explains why sequential processing can appear to be faster than parallel processing in the case of queries like these. What appears to be parallel processing corresponds to sequential processing plus the preceding phase of parallel processing up to the termination. Recommendation: Please check whether parallel processing is deactivated. Otherwise, consider deactivating parallel processing for those queries that are MultiProvider queries and have large result sets. Furthermore, starting with BW 3.0B Support Package 14 (BW 3.1 Support Package 8), you can set parameters to define the size of the maximum permissible interim overall result, that is, you can change the default valid value of 30,000 rows. For further details, please refer to SAP Notes 622841, 607164, and 630500. Please note that it is generally recommended that the number of base InfoProviders of a MultiProvider is not more than 10. A larger number of base InfoProviders is likely to result in a scenario where there are a lot more base InfoProviders than there are available dialog processes, resulting in limited parallel processing and many pipelined subqueries. Details can be found in SAP Note 629541. 5.7.1 Queries with more than 30000 rows accessing a MultiProvider Query name ZYM13_INV_Q0008 # Query Total DB exec. runtime [s] Time [s] 4 213.30 25.05 EarlyWatch Alert Report , , 8.03.2007 OLAP Time [s] 183.69 Frontend Time [s] 0.21 # Frontend # cells QDBTRANS 0 622553 10 BW Runtime Statistics Query name # Query Total DB exec. runtime [s] Time [s] 3 29.38 13.86 14 493.55 468.70 1 62.20 57.55 1 27.02 24.72 ZYM13_INV_Q0015 ZYMPAPL_1_Q0001 ZYM_PTFINV_Q0001 ZYM_PTFINV_Q0002 OLAP Time [s] 15.08 22.47 3.60 1.86 Frontend Time [s] 0.29 0.11 1.02 0.24 # Frontend # cells QDBTRANS 0 152331 0 210597 0 33187 0 32585 The table below contains a list of all Multiproviders, the number of their Infoprovider, and the name of their queries. Query ZYM13_INV_Q0008 ZYM13_INV_Q0015 ZYMPAPL_1_Q0001 ZYM_PTFINV_Q0001 ZYM_PTFINV_Q0002 MultiProvider YM13_INV YM13_INV YMPAPL_1 YM_PTFINV YM_PTFINV # InfoProvider 10 10 3 9 9 5.8 Upload of transactional data This section provides an overview of the upload of transactional data from 19.03.2007 to 26.03.2007. Please note that 'Time Target' can be bigger than the total time. This is because when the data is loaded to the target, the insert can be done in parallel. Total Total # of Requests 2 Total time [h] 0.0 Time Source [h] 0.0 Time Transfer [h] 0.0 Time PSA [h] 0.0 Time Update [h] 0.0 Time Target [h] 0.0 # records inserted 894 In the tables below, you can see an overview of the runtime in special areas for the TOP 5 requests in the last 7 days (19.03.2007 - 26.03.2007). SUMMARY Info-Source 0BWTC_C02 0BWTC_C05 # # Targets Requests 1 1 Records written 1 1 Time Time Source Transfer [sec] Rules [sec] 452 442 10.1 3.4 TOP 5 REQUESTS WITH HIGH RUNTIME IN UPDATE RULES Request- Target Info-Source Records ID written 65857 65856 0BWTC_C02 0BWTC_C02 0BWTC_C05 0BWTC_C05 452 442 Time Source [sec] 10.1 3.4 TOP 5 REQUESTS WITH HIGH RUNTIME IN SOURCE SYSTEM Request- Target Info-Source Records Time ID written Source [sec] 65857 65856 0BWTC_C02 0BWTC_C02 0BWTC_C05 0BWTC_C05 452 442 TOP 5 REQUESTS WITH HIGH RUNTIME IN DATA TARGET Request- Target Info-Source Records ID written 65857 65856 0BWTC_C02 0BWTC_C02 0BWTC_C05 0BWTC_C05 452 442 10.1 3.4 Time Source [sec] 10.1 3.4 0.2 0.0 Time PSA [sec] 0.0 0.0 Time Transfer Rules [sec] 0.2 0.0 Time PSA [sec] Time Transfer Rules [sec] 0.2 0.0 Time PSA [sec] Time Transfer Rules [sec] 0.2 0.0 Time PSA [sec] 0.0 0.0 0.0 0.0 0.0 0.0 Time Update Rules [sec] 0.3 0.1 Time Target [sec] Time Total [sec] 22.0 4.6 37.6 12.8 Time Update Rules [sec] 0.3 0.1 Time Target [sec] Time Total [sec] 22.0 4.6 37.6 12.8 Time Update Rules [sec] 0.3 0.1 Time Target [sec] Time Total [sec] 22.0 4.6 37.6 12.8 Time Update Rules [sec] 0.3 0.1 Time Target [sec] Time Total [sec] 22.0 4.6 37.6 12.8 5.8.1 Number Range Buffering for BW Objects For each characteristic and dimension, BW uses a number range to uniquely identify a value (SIDs and DIM-IDs). If the system creates a high amount of new IDs periodically, the performance of a data load may decrease. EarlyWatch Alert Report , , 8.03.2007 11 BW Runtime Statistics To avoid the high number of accesses to the NRIV table, activate the number range buffering for these BW objects (Main Memory Number Range Buffering). For more detailed information, see SAP Notes 504875, 141497, and 179224. The tables provide an overview of the number range buffering, setting of dimensions, and InfoObjects based on the highest number range level ("Level"). The purpose of this information is to identify the candidates for activating the number range main memory buffer. Recommendation: Activate number range buffering for all dimensions and InfoObjects, with a high number of rows, based on the rules mentioned in Note 857998. Note: A package dimension must not be buffered! TOP10 UNBUFFERED DIMENSIONS WITH HIGHEST NUMBER RANGE LEVEL InfoCube Dimension Rows Object 0BWTC_C02 0BWTC_C024 629238 BID0001142 100025 1000252 8495 BID0004427 100026 1000263 9784 BID0004434 100111 1001112 2835 BID0005162 0BWTC_C02 0BWTC_C021 123526 BID0001141 100004 1000042 7927 BID0004494 100147 1001472 7502 BID0005669 100027 1000273 1842 BID0004440 100025 1000253 1842 BID0004428 100026 1000262 626 BID0004433 TOP10 BUFFERED DIMENSIONS WITH HIGHEST NUMBER RANGE LEVEL InfoCube Dimension Rows Object ZCORDINV ZCORDINV3 1 BID0001311 YC12_DEL YC12_DEL2 49780 BID0000746 YC13_INV YC13_INV2 51090 BID0000775 YC_OO YC_OO3 1 BID0000730 YC_OO YC_OO2 1 BID0000729 ZCORDINV ZCORDINV2 1 BID0001310 ZCHSTLI ZCHSTLI2 1 BID0001370 YC12_DEL YC12_DEL3 24551 BID0000747 YC13_AGR YC13_AGR2 36620 BID0000766 ZCHSTLI ZCHSTLI3 1 BID0001371 Level 674193 206000 177000 134000 132486 127000 104000 99000 99000 97000 Level 252481 181286 176399 141937 125860 120108 109527 103245 98250 88330 TOP10 UNBUFFERED INFOOBJECTS WITH HIGHEST NUMBER RANGE LEVEL InfoObject SID Table Rows Object YDOC_NUM1 /BIC/SYDOC_NUM1 10769585 BIM0000802 0AC_DOC_NO /BI0/SAC_DOC_NO 2838940 BIM0000963 YPAOBJNR /BIC/SYPAOBJNR 4211590 BIM0000804 0REFER_DOC /BI0/SREFER_DOC 2484486 BIM0000533 0DOC_NUMBER /BI0/SDOC_NUMBER 2630186 BIM0000362 0REF_DOC_NO /BI0/SREF_DOC_NO 1327628 BIM0000908 0DOC_NUM /BI0/SDOC_NUM 1723056 BIM0000361 0DELIV_NUMB /BI0/SDELIV_NUMB 1120222 BIM0000345 0BILL_NUM /BI0/SBILL_NUM 1551025 BIM0000714 YRBELN /BIC/SYRBELN 1214712 BIM0000806 TOP10 BUFFERED INFOOBJECTS WITH HIGHEST NUMBER RANGE LEVEL InfoObject SID Table Rows Object 0MAT_SALES /BI0/SMAT_SALES 152286 BIM0000437 YSALFORCE /BIC/SYSALFORCE 116770 BIM0000723 0MAT_PLANT /BI0/SMAT_PLANT 91450 BIM0000942 0CUST_SALES /BI0/SCUST_SALES 64504 BIM0000336 0MATERIAL /BI0/SMATERIAL 49762 BIM0000426 0CUSTOMER /BI0/SCUSTOMER 19576 BIM0000325 0GL_ACCOUNT /BI0/SGL_ACCOUNT 10330 BIM0000985 ZFORMCODE /BIC/SZFORMCODE 5761 BIM0001177 0COSTELMNT /BI0/SCOSTELMNT 2502 BIM0000215 0COORDER /BI0/SCOORDER 767 BIM0000212 Buffered Buffered Level Buffered 12270434 5373185 4783483 3047153 2713585 1940988 1729291 1613569 1551355 1481971 Level 201646 120506 95285 68268 57121 20392 11107 6747 3071 1467 Buffered Buffer Size 0 0 0 0 0 0 0 0 0 0 Buffer Size 100 100 100 100 100 100 100 100 100 100 Buffer Size 0 0 0 0 0 0 0 0 0 0 Buffer Size 100 100 100 100 100 100 100 100 100 100 5.9 Process Chains - Runtime Overview EarlyWatch Alert Report , , 8.03.2007 12 BW Runtime Statistics The Process Chain Runtime Analysis is based on the last 7 days before the download. The table contains statistical information of all chains which aren't started by another local process chain. This includes process chains which are started by the Service API Interface or remotely by a chain from another system. Please note that only the top 20 chains having the longest runtimes are displayed. The columns '# Total Subchains', '# Total Steps' represent the summarized values of the main chain and their subchains. The runtimes are the range from the start of the main chain up to the end of the last process type executed within the main chain and their subchains. This means the real runtime of the main chain and their subchains is displayed here. Main Chain #Total Subchains ZBWD_SD_DM ZBWD_SD ZBWW_HSE_KEEP ZBWD_SD_V3 ZBW_STATISTICS_TRANS ZBWD_FI_DM ZTEMP_APD_SD_CARYOVR ZBWD_CR #Total #Runs Steps 7 6 10 6 9 5 0 0 44 38 91 17 105 30 4 3 14 14 1 14 7 7 2 7 Total Runtime [min] 1809 457 273 203 90 69 41 4 Avg. Runtime [min] 130 33 273 15 13 10 21 1 Med. Runtime [min] 104 14 273 11 13 11 21 1 Avg. Proc.Type Runt. [min] 130 33 314 29 13 10 21 1 The following table displays the statistical information based on the actual defined chain hierarchy. The columns '# Subchains' and '# Steps' are based on the subchain itself only. Only the top three main process chains with the highest number of executions are displayed. For each main process chains only the first ten subchains with the highest runtimes are shown. Parent Chain ZBWD_SD_DM ZBWD_DM_SD_DELTA_LOAD ZBWD_DM_SD_DELTA_LOAD ZBWD_DM_SD_DELTA_LOAD ZBWD_DM_SD_DELTA_LOAD ZBWD_DM_SD_DELTA_LOAD ZBWD_DM_SD_DELTA_LOAD ZBWD_SD ZBWD_SD ZBWD_SD ZBWD_SD ZBWD_DELTA_LOAD ZBWD_DELTA_LOAD ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP ZBWW_HSE_KEEP Level Chain 0 1 2 2 2 2 2 2 0 1 1 1 1 2 2 0 1 1 1 1 1 1 1 1 1 ZBWD_SD_DM ZBWD_DM_SD_DELTA_LOAD ZBWD_DM_SD_OO ZBWD_DM_SALES_ORD ZBWD_DM_SLS_INV_ROLLUP ZBWD_DM_SALES_INVOICE ZBWD_DM_SALES_DELIV ZBWD_DM_SD_SRI ZBWD_SD ZBWD_DELTA_LOAD ZBWD_RFC_LIS_APPL_11 ZBWD_RFC_LIS_APPL_12 ZBWD_RDC_LIS_APPL_13 ZBWD_SD_TRANS_LOAD ZBWD_SD_INV_TRANS_LOAD ZBWW_HSE_KEEP ZBW_DBSTATS_SD ZBW_ZCORDINV_COMPRESS ZBW_DBSTATS_FI ZBW_IDX_CREATE_SD ZBW_IDX_DROP_SD ZBW_COMPRESS_FI ZSD_PSA_DELETE ZBW_IDX_CREATE_FI ZBW_PSA_DELETE_FI # # # Total Avg. Me Subchains Steps Runs Runtime Runtime Runtim [min] [min] [mi 1 6 0 0 0 0 0 0 4 2 0 0 0 0 0 10 0 0 0 0 0 0 0 0 0 2 7 3 10 4 5 8 5 10 3 2 2 2 14 5 11 9 9 3 9 9 3 19 4 11 14 14 14 14 14 14 14 14 14 14 28 28 28 14 14 1 1 1 1 1 1 1 1 1 1 1809 1808 1109 553 142 142 115 110 457 425 121 113 20 415 60 273 164 98 32 7 3 3 2 1 1 5.10 Change-Run Analysis Total maintenance time of all change-runs within one week. Week 12.2007 11.2007 10.2007 09.2007 08.2007 EarlyWatch Alert Report , , 8.03.2007 # Change-Runs 19 17 18 19 18 Total Runtime [min] 345 213 391 407 376 13 130 130 80 40 11 11 9 8 33 31 5 5 1 30 5 273 164 98 32 7 3 3 2 1 1 1 1 2 1 BW Runtime Statistics Week 07.2007 06.2007 05.2007 04.2007 03.2007 EarlyWatch Alert Report , , 8.03.2007 # Change-Runs 18 14 19 19 18 Total Runtime [min] 261 1246 165 234 341 14 Workload Distribution 6 Workload Distribution The performance of your system was analyzed with respect to workload distribution. We did not detect any major problems that could affect the performance of your SAP System. Rating Check Workload by Application Module DB Load Profile 6.1 Workload by Application Module The following graphics or tables show how each application module contributes to the total system workload. Two workload aspects are shown: - CPU time: Total CPU load on all servers in the system landscape - Database time: Total database load generated by the application The #Customer# section contains all programs and transactions that start with Y or Z. The #Others# section contains all programs that cannot be assigned to an SAP standard application module (for example, an industry solution or a third party add-on development). Module BC Others BC-MID BW EarlyWatch Alert Report , , 8.03.2007 CPU Time in % 52.0 30.3 13.7 3.9 15 Workload Distribution Module Others BC BC-MID BW DB Time in % 39.9 39.8 15.6 4.7 6.2 DB Load Profile The following table or diagram shows the DB load caused by Dialog, RFC, and Background tasks over different time frames. The data given in the table represents the average number of database processes occupied by each task type in the database during the given time frames. These statistics are calculated as a weekly average, which means the average values over six working days with a unit of one hour. Periods between 00:00-06:00 and 21:00-24:00 contain an average value per hour. By comparing the load profiles for dialog and background activity, you have an overview of the amount of background activity during online working hours. EarlyWatch Alert Report , , 8.03.2007 16 SAP System Operating 7 SAP System Operating Your system was analyzed with respect to daily operation. We detected some problems that may impair system operation and stability. To ensure system stability, you should take corrective action as soon as possible. Rating Check System Availability Update Errors Table Reorganization ABAP Dumps 7.1 Update Errors In a system running under normal conditions, only a small number of update errors should occur. To set the rating for this check, the number of active users is also taken into consideration. We did not detect any problems. 7.2 Transports The following graphics contain information about the number of objects which were transported to the SAP system. The three graphics/tables show objects in total, programs and table (content). EarlyWatch Alert Report , , 8.03.2007 17 SAP System Operating 7.3 Program Errors (ABAP Dumps) 10 ABAP dumps have been recorded in your system since 21.03.2007. ABAP dumps are deleted after 7 days by default. To view the ABAP dumps in your system, call transaction ST22 and choose “Selection”. Then select a timeframe. ABAP DUMPS PER DAY Date 03/20/2007 03/21/2007 03/22/2007 03/23/2007 03/24/2007 03/25/2007 DETECTED ABAP DUMPS Error Number Server (e.g.) TSV_TNEW_BLOCKS_NO_ROLL_MEMORY 4 usnasap123 UNCAUGHT_EXCEPTION 4 usnasap123 SYSTEM_CORE_DUMPED 1 usnasap101 TIME_OUT 1 usnasap123 EarlyWatch Alert Report , , 8.03.2007 Quantity of Dumps 0 4 4 1 1 0 User (e.g.) US1592172 US1599967 US2WBAIG0 ALEREMOTE Date (e.g.) Time (e.g.) 03/21/2007 14:18:47 03/22/2007 19:31:42 03/23/2007 16:52:48 03/24/2007 19:31:09 18 Security Checks 8 Security Checks Critical security issues were detected in your system. We determined that several users have critical authorizations for certain transactions. See the information in the following checks. SAP offers a complete Security Optimization service for your system, which covers all of R/3 Basis and other SAP products. We recommend that you make use of this service. For more information, see SAP Service Marketplace at http://www.service.sap.com/SOS. Rating Check Users Authorized to Display all Tables Users Authorized to Start all Reports Users Authorized to Debug / Replace Users Authorized to Display Other Users Spool Request Users Authorized to Administer RFC Connections Users Authorized to Reset/Change User Passwords Password of One or More Standard Users is not Changed 8.1 Password of One or More Standard Users is not Changed Evaluated Risk - High Recommendation: Change the standard password for users SAP* DDIC, SAPCPIC, and EARLYWATCH in all clients. Do not delete the SAP* user because this enables the use of the hard-coded SAP* user. For more information, see the SAP Security Guides on SAP Service Marketplace at: http://www.service.sap.com/securityguide. Client 000 320 066 User SAPCPIC SAPCPIC EARLYWATCH EarlyWatch Alert Report , , 8.03.2007 19 Hardware Capacity 9 Hardware Capacity We have checked your system for potential CPU or memory bottlenecks and found that the hardware of your database and application servers is sufficient for the current workload. Server usnasap101 usnasap123 Max. CPU load Date Rating [%] 18 03/19/2007 74 03/25/2007 EarlyWatch Alert Report , , 8.03.2007 RAM [MB] 8192 18432 Max. Paging [% of Date Rating RAM] 0 03/21/2007 0 03/21/2007 20 Database Settings 10 Database Settings We have detected some problems with the settings of the database. These settings may affect performance. Rating Check Database Key Performance Indicators Locally managed Temp Tablespace Database Parameters Optimizer Statistics 10.1 Load per user The following table provides an overview of the load caused by different database users. Listed for each database user is the: (a) Percentage of executed statements (b) Number of block reads from disk (indicative for the I/O load caused by this database user) (c) Number of buffers accessed (d) Number of records returned as obtained from the shared pool (Oracle view V$SQL) at the time of the download The first line with the name TOTAL indicates the total number of (a), (b), (c) and (d) for all the user names. 10.1.1 User name Executions Disk reads Buffer gets Rows processed TOTAL (not a user name!) 1882491299 55777616 9947217155 5093278823 SYS OPS$ORAPWH OPS$PWHADM SAPBIW THEGUARD 2% 0% 0% 98% 0% 1% 0% 0% 27% 72% 13% 0% 0% 65% 23% 9% 0% 0% 73% 17% 10.2 File System Response Times Your file system read response times do not meet the standard requirements. Recommendation: Check the disks where the data files named in the following table are located. All datafiles with a high average read time and a noticeable impact on the total read time of your system are listed in this table. DATAFILES WITH HIGH AVERAGE READ TIME PER BLOCK File name Blocks read /oracle/PWH/sapdata1/biw_49/biw.data49 50196646 /oracle/PWH/sapdata2/biw_39/biw.data39 78018544 /oracle/PWH/sapdata3/biw_43/biw.data43 77675545 /oracle/PWH/sapdata4/biw_30/biw.data30 75837626 /oracle/PWH/sapdata1/biw_12/biw.data12 98598203 /oracle/PWH/sapdata2/biw_2/biw.data2 102701910 /oracle/PWH/sapdata2/biw_1/biw.data1 140915515 Avg. read time (ms) 264 191 184 174 139 124 108 Total read time per datafile (ms) 13251914544 14901541904 14292300280 13195746924 13705150217 12735036840 15218875620 10.3 Database Parameters We noticed that the following database parameters, which are currently active in your system, do not follow our standard recommendations. These standard recommendations cannot take into account possible special aspects for your system. Deviations from those standard recommendations may be necessary in your case. We therefore advise you to discuss theses parameter settings in your next Service session. EarlyWatch Alert Report , , 8.03.2007 21 Database Administration Based on the fact that you intend to use OLAP operations in your NetWeaver system's database, the recommended parameter settings are optimized for OLAP operations as listed in SAP Note 632556. While these settings should also provide adequate performance for OLTP database operations for most database accesses, some accesses may yield better results with the parameter settings listed in SAP Note 124361. This means that, depending on the priorities you assign to the OLTP or OLAP operations in your system (which we are not able to ascertain from the technical data available in a remote service session), the parameter settings in SAP Note 124361 may be preferable. In this case, ignore the following parameter recommendations. NOTE FOR ORACLE PARAMETER SETTINGS Note Description number 332696 ORA-00600 [2865], star_transformation_enabled 632556 Oracle 9.2.0.* database parameter settings for BW Parameters Description parallel_max_servers Parallel Query: Maximum number of parallel execution servers Maximum number of database shadow/background processes Maximum number of concurrent database sessions processes sessions Current value 48 270 330 Recommended value Please check SAP note 0632556 Please check SAP note 0632556 Please check SAP note 0632556 To provide special functionality or control system behavior, Oracle has a number of events. These can be activated in the configuration file and/or dynamically in the server manager or sqlplus. In general, these events and parameters should only be activated on request. When they are no longer needed, they should be removed. However, there might be the need to activate some events during normal database operation. The following table displays the values for the parameter "event" as found in your system: Parameter event Current Value 10028 trace name context forever, level 1 10191 trace name context forever, level 1 38060 trace name context forever, level 1 38069 trace name context forever, level 1 The following table lists events that may have to be activated. Please check the mentioned SAP Notes for details. Event Note number for Event Description 10027 596420 System downtime during deadlock (ORA-60) 38068 176754 Problems with CBO and RBO The following table lists events we found to be activated in your system, but which should only be activated in special circumstances. Please check whether there are special reasons to set these events. If they are no longer needed, they should be removed. Event 38060 11 Database Administration In the checks performed, problems regarding the administration of your database were found. Rating Check Database Growth Backup Frequency Archive Frequency Freespace in Tablespaces Space Critical Objects brconnect -f check (sapdba -check) schedule EarlyWatch Alert Report , , 8.03.2007 22 Database Administration 11.1 Database Growth The following figure shows the development of the size of your database in GB. An overview of the freespace development of your database in GB is shown here. The following table shows you the current size and the monthly growth of your database in GB. Date Current Size in GB 04/18/2006 13.38 05/31/2006 13.18 06/30/2006 100.72 07/31/2006 200.87 08/31/2006 256.84 09/30/2006 475.59 10/31/2006 570.98 11/30/2006 765.99 12/31/2006 560.60 01/31/2007 471.97 02/28/2007 353.34 03/25/2007 355.52 Your database is growing at a rate of more than 30 GB of data per month. Monthly Growth in GB 0.60 -0.20 87.54 100.15 55.97 218.75 95.39 195.01 -205.39 -88.63 -118.63 Recommendation: Analyze the top growing tables and check that the table growth is really necessary for your business process. Set up a archiving project that allows you to archive data periodically to keep the database at a manageable size. Note: Data archiving is particularly important before SAP Basis Release upgrades, conversion, or migration projects. EarlyWatch Alert Report , , 8.03.2007 23 Database Administration 11.2 Oracle Release In general, we recommend that you install the most current Oracle patch set as soon as possible (see SAP Note 415900). Therefore, check SAP Note 539921 from time to time for the latest patch set for Oracle 9.2.0. Note that in addition to the patch set, further bug fixes might be available and recommended. See the following SAP Notes: For patch set 9.2.0.3 - 610445: Merge fix for Oracle 9.2.0.3 For patch set 9.2.0.4 - 695080: Merge fix for Oracle 9.2.0.4 For patch set 9.2.0.5 - 755629: KKO merge fix for Oracle 9.2.0.5 For patch set 9.2.0.6 - 834100: Merge fix for Oracle 9.2.0.6 For patch set 9.2.0.7 - 896903: Optimizer Merge fix for Oracle 9.2.0.7 11.3 Desupport of multibyte character sets as of Oracle 10g An important SAP Note is available concerning systems that run with an Oracle database and that make use of setting the database code page to one of the following multibyte character sets: JA16SJIS Japanese KO16KSC5601 Korean ZHS16CGB231280 Simplified Chinese ZHT16BIG5 Traditional Chinese This SAP Note is: Note 858869: Desupport of multibyte character sets as of Oracle 10g. (see also related note 695899) The database settings as shown above have been set by many Asian customers to allow third-party tools to connect to their databases. With Oracle 10 clients or kernel release 7.00, the settings above are no longer supported and a Unicode conversion of the systems prior to an upgrade becomes mandatory for these systems. 11.4 Important Notes for Oracle To optimize system performance and to prevent data loss or system downtime, implement the following SAP Notes as soon as possible. Note number 858869 871455 Description Desupport of multibyte character sets as of Oracle 10g Performance when accessing DBA_SEGMENTS EarlyWatch Alert Report , , 8.03.2007 24 Business Information Warehouse 12 Business Information Warehouse Severe problems were detected, that may have serious impact on your system's performance and stability. You should take corrective action immediately. Rating Check BW Application Checks Data Distribution Analysis of Infoproviders Analysis of Aggregates Source System Settings Recommendations for BW Frontend Servers Planning Performance 12.1 BW Application Checks for PWH The BW application was analyzed with respect to design and data distribution. We did not detect any major problems that could affect the performance of your system. 12.1.1 Analysis of OLAP Cache The OLAP Cache is used for duplicated storing of often used data, whereby this data can be accessed quickly. In the tables below, you can find information regarding the size and the usage of the OLAP Cache. Cache usage of queries Defined Queries The OLAP Cache can buffer results from queries and can provide them again for different users and similar queries (that is, same queries or real subsets of them). The OLAP Cache reduces the workload of the DB server. The OLAP Cache can store the query results with their navigation statuses in the memory of the application server; the data can also be stored in database tables and files. When the buffer (Export/Import Shared Memory) memory overruns, it stores the displaced data - depending on the persistence mode - on the database server. The following OLAP Cache modes exist: 0 Cache is Inactive 1 Main Memory Cache Without Swapping 2 Main Memory Cache With Swapping 3 Cluster / Flat File Cache for Each Application Server 4 Cluster / Flat File Cache Across Each Application Server What is the best cache mode? This greatly depends on the requirements. The fastest cache mode is always the usage of the main memory. However, if several application servers are in use, the cache has to be filled for each query on each application server. In this case, it might be better to use the cross-application-server cache mode (flat file/cluster table). An alternative might be to use logon groups and to ensure that users executing the same/similar queries work on the same application server. MODE 0 - Cache is Inactive All data is read from the relevant InfoProvider. MODE 1 - Main Memory Cache without Swapping New data is stored in the export/import SHM buffer until this memory area is full. If then new data has to be added to the buffer, an LRU mechanism is applied. Least recently used data is DISPLACED from the buffer and lost. If this data is needed again by a query, it must access the relevant InfoProvider on the DB server. MODE 2 - Main Memory Cache with Swapping This works in a similar fashion to MODE 1. HOWEVER in this scenario if the memory is full, data is not DISPLACED and lost but written to a cluster table/flat file. Then if this data is needed again by a query, it can be read from this file, which is still quicker than reading this from the relevant InfoProvider on the DB server. EarlyWatch Alert Report , , 8.03.2007 25 Business Information Warehouse MODE 3 - Cluster/Flat File Cache for Each Application Server The cache data is kept persistently in a cluster table or in flat files for each application server. The overall data quantity is only restricted by the database or file system. Swapping does not occur in the same way as with the cache main memory mode. MODE 4 - Cluster/Flat File Cache Application Server-Extensive This mode is the same as the mode described above (cluster/flat file for each application server), the only difference being that the cache entries of all of the application servers in a system are used together. NOTE: FOR MODE 1 and 2, remember that MEMORY is INSTANCE DEPENDENT. NUMBER OF QUERIES PER CACHE MODE Cache Mode Total OFF 1 2 3 4 # Queries 180 0 146 34 0 0 NUMBER OF INFOCUBES PER CACHE MODE Cache Mode Total OFF 1 2 3 4 # Infocubes 69 0 34 35 0 0 Executed Queries Task type Nav. Steps Accessed DB [%] Accessed Cache [%] RSDRI Queries [%] All Queries 6676 61 39 0 There are two types of caches: The local cache and the transactional cache (OLAP Cache). The local cache belongs to a query session and therefore cannot be used by other sessions. The OLAP Cache can store query data on the application server and can have a swap file or use a swap cluster table. The OLAP Memory Cache is located in the Export/Import buffer SHM (parameter rsdb/esm/buffersize_kb). As the global cache size is a logical value and the Export/Import SHM gives a physical limit, we recommend that you do not set the global cache parameter to a higher value than the Export/Import SHM buffer. Note: The OLAP Cache was optimized in SAP BW 3.0B SP19, SAP BW 3.1C SP13, and SAP BW 3.5 SP02. For more information, see SAP Note 683194. Rating Description Cache active Cache Persistence Mode Flat File Name EarlyWatch Alert Report , , 8.03.2007 Current Value Active Flat File BW_OLAP_CACHE Recommendation 26 Business Information Warehouse Rating Description Comprehensive Flat File Name for AppServer Local Cache Size (MB) Global Cache Size (MB) Current Value Recommendation 100 200 12.2 Data Distribution 12.2.1 Largest ODS tables ODS Name ZOHSTLII YOPAPL_2 YO11_SCL YO_OO_I YO11_ITM YO12_ITM YO11_SRI 0FIGL_O02 YOPAOC_1 YO13_ITM ODS Table name /BIC/AZOHSTLII00 /BIC/AYOPAPL_200 /BIC/AYO11_SCL00 /BIC/AYO_OO_I00 /BIC/AYO11_ITM00 /BIC/AYO12_ITM00 /BIC/AYO11_SRI00 /BI0/AFIGL_O0200 /BIC/AYOPAOC_100 /BIC/AYO13_ITM00 No of records 34024444 30616050 16877028 14338252 13498023 11057837 10741688 10697281 10107754 8661444 Large ODS Objects can have a negative impact on reporting and upload performance. Please pay attention to more detailed recommendations in subsequent sections of this report. With BW release 3.x parallel ODS Data Activation can fail on large ODS Objects with a deadlock on the Oracle database. Recommendation: Please follow the implementation description in SAP Note 634458 "ODS object: Activation fails - DEADLOCK" and 84348 "Oracle deadlocks, ORA-00060". 12.2.2 Largest InfoCubes InfoCube Name YC11_ORD YCPAPL_1 ZC_OO ZCINVL05 ZCINVL04 ZCORD_L06 ZCODI_L06 ZCORD_L07 ZCODI_L07 ZCPAPLL07 No of records 23743474 16069421 15215471 13741030 12958370 11325095 10223447 10196204 9579417 9161759 Large InfoCubes require high effort in administration and have a long runtime during maintenance activities. Recommendation: Please consider splitting a large InfoCube into small InfoCubes (such as separated by a time characteristic like 0CALYEAR). You can use MultiProviders as query targets to obtain all data in your query result. 12.2.3 Largest Aggregates Infocube ZCINVL04 ZCINVL05 ZCINVL06 ZCINVL05 ZCINVL04 ZCODI_L07 ZCODI_L06 ZC_OO ZCODI_L07 ZC_OO Aggregate Name 100123 100125 100126 100103 100097 100195 100139 100155 100200 100147 No of records 8946973 8905517 4404831 3533970 3531850 3512065 3284100 2460809 2204225 2188598 Large aggregates need high runtime for maintenance like change runs and rollup of new data. Recommendation: Please check section: "Analysis of Aggregates" for further hints. EarlyWatch Alert Report , , 8.03.2007 27 Business Information Warehouse 12.2.4 Largest Master data tables (SID-tables) Master data SID-Table YDOC_NUM1 YPAOBJNR 0AC_DOC_NO 0DOC_NUMBER 0REFER_DOC Table Name /BIC/SYDOC_NUM1 /BIC/SYPAOBJNR /BI0/SAC_DOC_NO /BI0/SDOC_NUMBER /BI0/SREFER_DOC No of records 10769585 4211590 2838940 2630186 2484486 The table contains a list of the largest SID tables in your system. Recommendation: Please ensure that you have switched on number range buffering for all InfoObjects where a high number of tuples are loaded. This is described in SAP Note 130253 "Notes on upload of transaction data into the BW." 12.2.5 Largest Master data tables (time independent: X-tables) Master data X-Table 0DOC_NUMBER 0MAT_SALES YSALFORCE 0MAT_PLANT 0CUST_SALES Table Name /BI0/XDOC_NUMBER /BI0/XMAT_SALES /BIC/XYSALFORCE /BI0/XMAT_PLANT /BI0/XCUST_SALES No of records 2198774 152285 116770 89530 64504 The table above shows the largest tables with time-independent navigational attributes. The usage of navigational attributes in aggregates will increase the runtime of the change run. 12.2.6 Largest Master data tables (time dependent: Y-tables) Master data Y-Table 0PERSON 0COSTCENTER 0ORGUNIT 0EMPLOYEE 0HRPOSITION Table Name /BI0/YPERSON /BI0/YCOSTCENTER /BI0/YORGUNIT /BI0/YEMPLOYEE /BI0/YHRPOSITION No of records 1687 813 1 1 1 The table above shows the largest tables with time-dependent navigational attributes. Time-dependent navigational attributes can have a negative impact in the runtime of queries. Recommendation: Please consider the usage of such objects carefully. 12.2.7 Largest Hierarchy tables (I-tables) Master data Hierarchy 0MAT_SALES 0PROD_HIER YPRODH2 0REASON_CDE 0BUS_AREA Table Name /BI0/IMAT_SALES /BI0/IPROD_HIER /BIC/IYPRODH2 /BI0/IREASON_CDE /BI0/IBUS_AREA No of records 356158 44104 577 172 122 Large hierarchy I-Tables will have a negative impact on the runtime of queries using those hierarchies. Recommendation: Please see SAP Note: 738098 "Performance problems with hierarchies" 12.3 Analysis of InfoProviders 12.3.1 InfoProvider Distribution The following section provides an overview of the distribution of your InfoProviders. Note that the following overview table takes into account only objects that can currently be used for reporting. InfoProviders BasisCubes MultiProviders Aggregates Virtual Remote Transactional ODS Info InfoSets Cubes Cubes Cubes Objects Objects 327 46 23 153 0 0 0 23 78 4 Aggregates The table below displays the top 3 InfoCubes regarding the number of their aggregates. InfoCube ZCODI_L07 ZCODI_L06 EarlyWatch Alert Report , , 8.03.2007 # Aggregates 9 9 28 Business Information Warehouse InfoCube ZCODI_A07 # Aggregates 9 ODS Objects The following table gives a summary of the ODS objects. Description: BEx Indicator With the BEX indicator, you can determine whether the ODS object is immediately available for BEX queries. If the ODS object is not used for reporting, but only as a data container, you can delete the BEx Reporting indicator. The advantage is that no SIDs are generated, which leads to a better activation performance. If your ODS object reporting requirements are very "simple," in other words, if you are only selecting a limited number of records, you can use InfoSets for reporting on ODS objects even if the BEx reporting indicator is switched off. Please note that the loaded values are not checked for their correctness if SIDs are not generated. If you update the data of the ODS object in other data targets (DataMart functionality), SIDs are generated during this update should the BEx reporting indicator not be set. Additionally, it is possible to set the BEx reporting indicator even if the ODS object already contains data. In this case, the system generates the SIDs. Unique Data Records If each key field combination is only loaded once, you can set the indicator 'Unique Data Records,' which leads to a performance enhancement during activation. The system then does not need to check the existence of a record in the active data table and therefore a mass insert into the active table can be executed directly (otherwise, checks are first carried out to establish whether the datasets are already available in the active data table and then the dataset is marked either for update or insert). Additionally the data is not sorted. Attention: If the indicator 'Unique Data Records' is activated, the BW system does not guarantee unique records by deleting duplicates with regard to the ODS object key fields. Therefore the DataSource has to make sure that unique records are always delivered. If this is not the case, the ODS data activation will terminate. # ODS Objects 40 # ODS Objects with BEX flag 23 # ODS Objects with unique flag 6 # Transactional ODS Objects 2 12.3.2 Dimensions with line item and high cardinality flag We checked for InfoCubes that contain dimensions with an active "High-Cardinality" indicator and determined that the current settings are the same as our recommended settings. Background: When this indicator is set, the database will create B*-tree indexes instead of bitmap indexes for this dimension. This can lead to a reduction in performance, as only bitmap indexes can be used for a Star-join access, which is the most effective way of selecting data. 12.3.3 InfoCube Design of Dimensions We checked for InfoCubes with one or more dimensions containing 30% or more entries compared to the number of records in the fact tables and found that your InfoCubes' design is according to our recommendations. Explanation: The ratio between the number of entries in the dimension tables and the number of entries in the fact table should be reasonable. If an InfoObject has almost as many distinct values as there are entries in the fact table, the dimension this InfoObject belongs to should be defined as a line item dimension. Instead of creating a dimension table that has almost as many entries as the fact table, the system then writes the data directly to the fact table. On the other hand, if there are several dimension tables with very few entries (for example, less than 10), those small dimensions should be combined in just one dimension. In order to obtain this information for your InfoCubes: - Call transaction RSRV. - Choose "All Elementary Tests" - "Database." - Double-click the line "Database Information about InfoProvider Tables." - In the window on the right, choose "Database Information about InfoProvider Tables." - Enter the InfoCube name and choose "Execute Tests." - After the analysis finishes, choose "Display Messages" and open the analysis tree with the correct time stamp. 12.3.4 InfoCube partitioning InfoCube ZCODI_A06 ZCODI_L06 Range partition used 0CALMONTH 0CALMONTH EarlyWatch Alert Report , , 8.03.2007 Partitioning Valid From 200607 200607 Partitioning Valid To 200612 200612 29 Business Information Warehouse We have determined that the partitioning validity has been exceeded for the InfoCubes in the table above. These InfoCubes are currently active and have a range partitioning validity date of last year or before. If you exceed the validity date for range partitioning, a single, new, and ever-growing partition will be created, which will contain all the latest data and all the previous data since the validity date was exceeded. This can lead to performance problems since the partition can grow quite large and will be accessed continually for the most recent information. It will contain all of the data added to the cube since the validity date was exceeded. Number of InfoCubes partitioned until this year 10 F-Fact table partitioning Table names /BIC/F100146 /BIC/F100147 /BI0/F0BWTC_C02 /BI0/F0BWTC_C05 /BI0/F0BWTC_C04 /BI0/F0BWTC_C03 /BIC/FYC11_ORD /BI0/F0BWTC_C09 /BIC/F100155 /BIC/F100156 /BI0/F0BWTC_C11 /BIC/F100104 /BIC/F100071 /BIC/F100102 /BIC/F100098 /BIC/F100096 /BIC/F100069 /BIC/F100103 /BIC/F100070 /BIC/F100097 /BIC/F100105 /BIC/F100072 /BIC/FZC_OO /BIC/F100099 /BIC/F100145 /BIC/F100144 /BIC/FZCORD_L07 /BIC/FZCORD_A07 /BIC/F100205 /BIC/F100207 /BIC/F100026 /BIC/F100025 /BIC/F100027 /BIC/F100202 /BIC/F100203 /BIC/F100204 /BIC/FZCSRIA07 /BIC/FYC13_AGR /BIC/F100077 /BIC/FZCINVL07 /BIC/F100091 /BIC/FZCSRIL07 /BIC/F100164 /BIC/FZCINVA07 /BIC/F100165 /BIC/F100125 /BIC/F100126 /BIC/F100123 /BIC/F100093 /BIC/F100085 /BIC/F100092 /BIC/FZCDELL07 EarlyWatch Alert Report , , 8.03.2007 No. of Partitions 286 286 232 231 225 223 197 174 147 116 107 96 94 88 86 85 83 79 76 75 61 60 59 59 59 59 54 53 52 52 50 50 49 49 49 49 45 45 44 43 43 43 41 41 39 39 38 38 37 36 36 34 30 Business Information Warehouse Table names /BIC/F100079 /BIC/F100167 No. of Partitions 34 32 The tables listed in the table 'F-Fact tables' have more than 30 partitions. Recommendation: To avoid performance issues when these tables are read, we recommend that you use compression on the InfoCubes of these F-Fact tables. Please see SAP Note 590370 for further details. Empty F-Fact table Partitions F-fact table /BI0/F0BWTC_C11 /BIC/FZCORD_L07 /BIC/FZCORD_A07 /BIC/FZCDELL07 /BIC/FZCINVA07 /BIC/F100062 /BIC/F100024 /BI0/F0BWTC_C09 /BI0/F0BWTC_C05 No.Empty Partitions 4 2 2 2 2 1 1 1 1 The #Empty Partitions# table contains up to ten F-fact tables with the highest number of empty partitions. There are 21 F-fact tables in total that contain empty partitions in your system. To check for tables with empty partitions, use report SAP_DROP_EMPTY_FPARTITIONS. See SAP Note 430486 for the appropriate recommendations for an overview or repair of F-fact tables in a BW InfoCube. Note: If the values shown in the table above do not seem reasonable when compared to report SAP_DROP_EMPTY_FPARTITIONS, the database statistics are probably not up to date and we were therefore not able to collect the most current data. To fix this, refresh the statistics. E-Fact-Table partitioning (Search for E Fact tables with many parts ) There is no action currently required as we have not found any E-Fact tables with 100 partitions or more. 12.4 Analysis of Aggregates Aggregates only improve performance when they are used by your queries and when they summarize the data of the structure they are built from (the InfoCube or another aggregate). Unused or incorrect aggregates consume space on your database and increase the time needed for the roll-up and the change run procedure. For this reason, you should create proper aggregates in your system and regularly check that you are using the proper aggregates. We offer the following training courses for performance optimization: TEWA50 SAP BW Query tuning with aggregates. For more details about training, refer to the Internet address http://service.sap.com/empoweringworkshops 12.4.1 Aggregates suggested for deactivation We have found aggregates that are not used during query execution ("Number of calls" is equal to zero), or that are almost as large as the parent aggregate they are built from (the "Average reduce factor" is too low). Recommendation: Consider deactivating or deleting the aggregates listed below. Note: Do not deactivate an aggregate that has been created to improve the performance of master data activation. This kind of aggregate (Basis aggregate) does not contain any navigational attributes or hierarchies. This means it is not affected during master data activation. Aggregates that must be recalculated during master data activation can access the Basis aggregate instead of the InfoCube during master data activation. Cube name YC12_DEL YC13_AGR YC13_AGR YC13_INV YC13_INV YC13_INV YCPAPL_1 ZCDELA07 ZCDELL07 ZCINVA04 ZCINVA05 Aggr.-cube 100020 100025 100111 100012 100106 100114 100028 100166 100167 100116 100117 # entries 147931 4247937 84588 1718210 1696534 51855 20376 32795 131987 100496 111371 EarlyWatch Alert Report , , 8.03.2007 Average reduce factor 79 2 1 2 1 1 1 20 102 1 1 # calls 0 1811 426 9638 94 68 1300 0 0 282 214 Created at 23.01.2007 10.09.2006 31.10.2006 18.01.2007 05.02.2007 05.02.2007 14.01.2007 23.01.2007 23.01.2007 31.10.2006 31.10.2006 Last Call ..0 25.03.2007 03.03.2007 26.03.2007 01.03.2007 03.03.2007 26.03.2007 ..0 ..0 03.03.2007 03.03.2007 31 Business Information Warehouse Cube name ZCINVA06 ZCINVA07 ZCINVA07 ZCINVL04 ZCINVL07 ZCINVL07 ZCINVL07 ZCINVL07 ZCODI_A06 ZCODI_A06 ZCODI_A06 ZCODI_A07 ZCODI_A07 ZCODI_A07 ZCODI_L06 ZCODI_L06 ZCODI_L06 ZCODI_L06 ZCODI_L07 ZCODI_L07 ZCODI_L07 ZCORD_A06 ZCORD_A07 ZCORD_L07 ZCPAPLA07 ZCPAPLA07 ZCPAPLL07 ZCPAPLL07 ZCPAPLL07 ZCPAPLL07 ZCSRIL07 ZC_OO Aggr.-cube 100118 100173 100171 100123 100176 100180 100179 100181 100129 100152 100132 100186 100191 100189 100139 100154 100159 100141 100195 100201 100198 100134 100204 100207 100221 100209 100220 100216 100217 100214 100165 100156 # entries 63685 28073 25844 8945325 2138688 1272107 1216995 0 377182 120022 38251 349526 145361 60849 13500750 7149655 1862139 699920 10351102 2415924 965680 123998 98119 923215 7990 1970 136819 13207 13207 10826 244813 3316409 Average reduce factor 1 1 1 2 2 2 1 0 1 1 1 2 1 1 1 2 1 1 2 1 1 2 1 2 47 1 66 694 1 1 14 1 # calls 214 164 212 44 872 5378 212 198 3401 68 20846 3222 42 16393 3401 8317 68 20785 3196 42 17226 111 102 102 0 188 0 0 154 182 0 68 Created at 31.10.2006 19.01.2007 19.01.2007 31.10.2006 19.01.2007 19.01.2007 19.01.2007 19.01.2007 10.12.2006 10.12.2006 10.12.2006 27.12.2006 27.12.2006 27.12.2006 10.12.2006 10.12.2006 10.12.2006 10.12.2006 27.12.2006 27.12.2006 27.12.2006 12.12.2006 27.12.2006 27.12.2006 23.03.2007 10.03.2007 23.03.2007 10.03.2007 10.03.2007 10.03.2007 02.01.2007 10.12.2006 Last Call 03.03.2007 03.03.2007 01.03.2007 06.03.2007 26.03.2007 26.03.2007 01.03.2007 03.03.2007 26.03.2007 23.03.2007 26.03.2007 26.03.2007 23.03.2007 26.03.2007 26.03.2007 26.03.2007 23.03.2007 26.03.2007 26.03.2007 23.03.2007 26.03.2007 13.03.2007 13.03.2007 13.03.2007 ..0 26.03.2007 ..0 ..0 23.03.2007 26.03.2007 ..0 23.03.2007 12.4.2 Aggregates Maintenance The tables below show an overview of maintenance costs for your aggregates in the last six weeks. The first table shows the sum of all aggregates for each InfoCube. The second table shows up to 20 aggregates that required the most total maintenance time. In addition, we analyzed the three aggregates with the most "total time" in more detail. CUBES [14.03.2007 05:42:03 26.03.2007 04:56:02] InfoCube Nr. of Total time % Rollup Aggregates [sec] time Total 52 16812.5 YC13_INV 8 8383.4 11 YC13_AGR 5 1851.4 86 ZCINVL07 8 1533.7 100 ZCINVA07 8 1223.1 100 ZCPAPLL07 6 943.7 100 ZCPAPLA07 6 787.4 100 YCPAPL_1 4 459.8 100 ZCDELL07 1 369.7 100 ZCSRIL07 1 306.1 100 ZCDELA07 1 282.1 100 ZCSRIA07 1 262.2 100 YCZO_1 1 171.0 13 YC12_DEL 1 147.3 88 YC11_SR 1 91.6 93 AGG. [14.03.2007 05:42:03 26.03.2007 04:56:02] InfoCube Aggregate Total time % Rollup [sec] time Total 16812.5 YC13_INV 100060 2361.6 8 EarlyWatch Alert Report , , 8.03.2007 % (Re)creation time % Delta Changerun time 89 0 0 0 0 0 0 0 0 0 0 87 0 0 0 14 0 0 0 0 0 0 0 0 0 0 12 7 % (Re)creation time % Delta Changerun time 92 0 Records inserted 13342801 12894367 61717 237149 14051 35746 6046 20 10673 23903 2261 1712 55156 0 0 Records inserted 13342801 4125936 32 Business Information Warehouse InfoCube Aggregate YC13_INV YC13_INV YC13_INV YC13_INV YC13_AGR YC13_AGR YC13_AGR YC13_INV ZCDELL07 ZCINVL07 ZCSRIL07 ZCDELA07 ZCSRIA07 YC13_INV YC13_AGR ZCINVL07 ZCPAPLL07 YC13_AGR ZCINVL07 100012 100106 100010 100013 100025 100026 100027 100114 100167 100176 100165 100166 100164 100014 100024 100180 100215 100111 100182 Total time [sec] 2166.1 1515.7 867.3 746.3 536.6 489.8 398.2 393.2 369.7 338.8 306.1 282.1 262.2 247.9 230.2 220.6 202.3 196.6 195.0 % Rollup time 8 11 11 13 78 76 100 19 100 100 100 100 100 30 100 100 100 83 100 % (Re)creation time 92 89 89 87 0 0 0 81 0 0 0 0 0 70 0 0 0 0 0 % Delta Changerun time 0 0 0 0 22 24 0 0 0 0 0 0 0 0 0 0 0 17 0 Records inserted 3436420 3393068 544713 953134 36206 16692 7257 103710 10673 74909 23903 2261 1712 337386 474 47355 16224 1088 23530 12.5 Source System Settings Upload customizing BW (SBIW) In every SAP source system, part of the ROIDOCPRMS table controls the data transfer from the source system to the system. The table contains the following information: MAXSIZE - Maximum size of a data package in KB STATFRQU - Number of packets that are transferred before statistical information is sent MAXLINES - Maximum number of records sent in one data package MAXPROCS - Maximum number of dialog work processes for each upload request used to send the data to the system To ensure the performance and stability of the upload process, it is important that this table is set up correctly. We have checked the setup of this table for all source systems connected to the system. Recommendation: To avoid potential extraction problems, change the entries in table ROIDOCPRMS in the different source systems as explained in the table below. Consequences: If the setting for MAXSIZE is too large, data packets will also be very large. A system timeout may occur for work process(es) on the server assigned to these packets, or there may be problems with the memory. If the setting for MAXLINES is too large, the high number of records for each data packet might lead to dumps in the source system or memory problems on the server. If the values are too small, several small data packets are created, which increases the upload time. Note that you do not have to reduce the values if you have enough resources available (memory, work processes) and your parameter settings are tested carefully. Implementation: To implement those settings you have two possibilities: a) Modify all source systems centrally using the Administration Workbench of Call transaction RSA1, click a source system using the secondary mouse button, and choose #Customizing for Extraction#. The system navigates directly to the selected source system in transaction SBIW. Choose General Settings -> Maintain Control Parameters for the Data Transfer. b) Maintain the settings in the different R/3 source systems. Call transaction SBIW and choose General Settings ->Maintain Control Parameters for the Data Transfer. Note: If table ROIDOCPRMS is empty, the systems use default values during runtime. The defaults are 10,000 for MAXSIZE and 100,000 for MAXLINES. The default for MAXSIZE is small whereas the default for MAXLINES is reasonable for most systems. In some cases, it may be necessary to set the values for some of the parameters outside our recommended range. One example of this is the MAXPROCS parameter (maximum number of dialog work processes for each upload request). This parameter limits the use of resources during an upload in order to prevent a resource bottleneck occurring in the system. If you reduce the setting for this parameter, the upload performance may suffer even though there are enough system resources available. If you use your as a DataMart, you have to maintain this table in the system as well. EarlyWatch Alert Report , , 8.03.2007 33 Business Information Warehouse System name P60CLNT320 P60CLNT320 P60CLNT320 Parameter name maxsize statfrqu maxprocs Current value 000000 00 00 Recommended value 20000 10 3 12.5.1 Important Notes for Source System Upload Settings To optimize upload performance and to prevent any loading problems for the R/3 source systems, check and implement the following SAP Notes as soon as possible. Note Number 327828 385741 351936 Description Performance of BW supply systems Collective runs are not deleted Problems with different collective runs 12.6 Recommendations for BW 12.6.1 General Parameters The table shown here provides an overview of the Customizing parameters of the table 'RSADMIN' of which some or all are considered INVALID. Each parameter is assigned a rating in the column 'Invalid,' which details the reason for the parameters validity. Additionally, a check is made as to whether a duplicate exists. Only parameters that are invalid/duplicates and/or have an OSS note available will be displayed. The validity rating is explained below, sequence 0-5 : Invalid 0 = This means the parameter is valid and used correctly. Invalid 1 = This parameter is invalid and you should reconsider the setting, because it is either not recognized as existing or your current release does not support the use of this parameter. Invalid 2 = This parameter is invalid and you should reconsider the setting, because a value is mandatory for this parameter but none was found. Invalid 3 = This parameter is invalid and you should reconsider the setting, because it is not relevant for your current database. Invalid 4 = This parameter is invalid and you should reconsider the setting, because your current support package does not support its use. Invalid 5 = This parameter is invalid and you should reconsider the setting, because your SAPI release is not sufficient. Recommendation: Please verify if the parameters showing as invalid can be removed or have to be changed. If duplicates exist, please remove the incorrect version. PARAM_NAME BW_USER CR_MAXWPC CR_RFCGROUP CSV_MIMETYPE IS_MO_AUTH_CHECK ORA_CUBEINDEX_INITRANS ORA_ODS_TABLE_INITRANS REUSE_XL_INSTANCE VALUE BIWREMOTE 6 PARALLEL_GENERATORS 20 20 OSS_NOTE 410952 534630 534630 705011 766432 750033 831234 709141 INVALID 0 0 0 2 2 0 0 2 DUPLICATE 12.6.2 BW Support Package The Support Package applied to your BW release 3.50 is not up-to-date. Recommendation: Apply the most recent Support Package. Implementation: For information about implementing SAP Support Packages, see SAP Note 97621. Also check the SAP Notes mentioned in the following tables. BW CURRENT SUPPORT PACKAGE Product Release Support Package Level 3.50 13 BW AVAILABLE SUPPORT PACKAGES Product Release 3.50 EarlyWatch Alert Report , , 8.03.2007 BI Content Add-on 353 Support Package Level 17 Add-on Support Package 0003 Notes 872277 34 Business Information Warehouse BW AVAILABLE CONTENT ADD-ONS BI Content Add-ons 353 702 703 Content Support packages 7 4 1 Notes 786286 908488 908176 Please also check if you need to update your applied Support Package Stack (SP Stack). SP Stacks are released quarterly as a set of support packages and patches offered for a product version and its underlying components. The contained support packages must be used in the given combination. We recommend regular application of these SP Stacks. More information can be found in service marketplace under Quick Link /sp-stacks. 12.6.3 Important notes for BW The table below lists important SAP Notes for BW that deal with performance and data security. Note Number 0309824 0166433 0423099 0558994 0550669 0567747 0567746 0567745 Description BW problem message / Support Packages, Patches Options to find aggregates "Low Speed Connection" Parameter and BEx Analyzer General information on BW front-end patches Compressed transfer of BW Web Applications Composite note BW 3.x performance: Extraction Composite note BW 3.x performance: Query Composite note BW 3.x performance: DB-specific DATA CONSISTENCY Note Number 598691 323140 529789 487393 534295 565086 Description Check: InfoObjects with ALPHA and fields without ALPHA DUPREC short dump when loading master data BW extraction/extractor checker differences Incorrect values with star_transformation_enabled=true Incorrect values in the BW environment Composite SAP note: Incorrect data in DataMart or Query Area Check Tools Data Loading Data Loading Database Database Query 12.6.4 BI - KPIs There are BW Key Performance Indicators that exceed their reference value. This indicates severe problems, or at least optimization potential for performance, data volumes, and administration. Based on these key figures, we recommend that you order the corresponding services for further analysis. The table below shows the KPIs, their observed value, and the reference limit. The last column contains the Service Code for finding the appropriate service in the second table, "Available Services". RECOMMENDATIONS KPI Description NrDelAggr (#) Nr of Aggregates recommended to delete AVAILABLE SERVICES Service Service Name Code G1 General link to support services SAFEGUARDING S1 EarlyWatch Service O1 O2 O3 O4 Observed Reference Rating OPTIMIZATION BPPO - Extraction and Loading BPPO - Query Tuning BPPO - Planning and Simulation SAP Remote Performance Optimization EarlyWatch Alert Report , , 8.03.2007 43 20 Description Recommended Services YELLOW E1 Link http://service.sap.com/supportservices Remote analysis of system parameters and technical performance tuning http://service.sap.com/safeguarding Optimization of critical load path http://service.sap.com/bppo Optimization of query execution Optimization of Planning and Simulation Application http://service.sap.com/bppo http://service.sap.com/bppo http://service.sap.com/rpo 35 Business Information Warehouse Service Code O5 O6 E1 E2 Service Name Description Link SAP System Administration for BI SAP Data Volume Management for SAP BW EMPOWERING TEWA50 - Query tuning with aggregates PDEBW2 - BI Reporting Authorizations Knowledge transfer for optimal business process administration Analysis of disk space saving potential http://service.sap.com/sysadmin Workshop - Efficient Aggregate Control and Query Tuning Workshop - Reporting Authorization Setup and Optimization Official classroom training - TEWA50 http://service.sap.com/dma Official classroom training - PDEBW2 12.6.5 Nametab inconsistencies NAMETAB INCONSISTENCIES Table DDNTT DBDIFF RSDD_TMPNM_ADM DBA_VIEWS # Total 1953 4629 697 0 # View 01 0 301 307 0 # View 02 47 47 47 0 # View 03 1566 1565 1 0 # View 04 0 1371 1 0 # View 05 0 1345 1 0 # View 06 340 0 340 0 There are several entries in tables DDNTT and DDNTF that cannot be found in tables DBDIFF and RSDD_TMPNM_ADM (or in the DB_VIEWS of Oracle). This means that these temporary entries are obsolete and no longer used. SAP Note: 449891 Temporary database objects in BW 3.x Actions: 1. If there are too many entries in DBDIFF: After a change run, obsolete entries may still appear in table DBDIFF. This table has buffer mode FULL => for example: one day it grew up to 247MB. The problem is solved in the following Support Packages: - BW 3.0B Support Package 22 - BW 3.10 Support Package 16 - BW 3.50 Support Package 4 Report SAP_UPDATE_DBDIFF clears these obsolete entries from table DBDIFF. The problem may occur also in systems containing the correct Support Package if the correction report has not yet been run. SAP Note: 736845 After Change Run entries in DBDIFF resist 2. If there are several entries for 03 views in DBDIFF: Use report SAP_UPDATE_DBDIFF to clean DBDIFF. The report checks temporary objects in table RSDD_TMPNM_ADM and deletes any unnecessary ones. 3. If there are several entries for 06 views in RSDD_TMPNM_ADM: SAP Note: 698280 Reduce number of 06-er Tables in the database 4. If there are too many entries for 01 views in RSDD_TMPNM_ADM: SAP Note: 811354 Usage of complex queries reduces the performance of the database 5. If there are too many entries for 03 views in RSDD_TMPNM_ADM: Use the function module RSDDTMPTAB_CLEANUP as described in SAP Note: 867000 Inconsistencies in temporary tables of TYPE 06 6. If there are too many entries for 03 views in DDNTT and DDNTF compared to the number of entries in DBDIFF/RSDD_TMPNM_ADM. Several obsolete entries in the dictionary tables that are not cleaned Use the SAP_DROP_TMPTABLES report as described in SAP Note: 883843 "BW instances can no longer be started" Final result: Run SAP_DROP_TMPTABLES, SAP_UPDATE_DBDIFF or RSDDTMPTAB_CLEANUP depending on the situation to clean obsolete temporary entries. 12.7 Frontend Servers EarlyWatch Alert Report , , 8.03.2007 36 Business Information Warehouse To ensure the best performance and stability on your frontend servers, you must install the most recent frontend server release and patch level. From frontend server 2.0B patch level 6 onwards, the installation data of the frontend server is written to the BW server table RSFEC. During this session, we downloaded and checked the most recent information (in a time frame of two weeks) with respect to the frontend server release and patch level used and the installation of a maximum of 100 frontend servers. Even though we checked up to 100 frontend servers, other frontend servers may still have severe problems. We cannot check frontend servers with an older release or patch level than frontend server 2.0B patch level 6. The following table gives an overview of the results. SUMMARY OF CHECKED PCS Rating Number of PCs checked Green 0 Yellow 0 Red 6 For more information about the problems we detected, please refer to the following sections. 12.7.1 PCs with outdated Patch Level The frontend patch installed on some of your frontend servers is not current. Recommendation: Immediately update the frontend servers listed in the following table to the newest frontend patch available. PC Name USCORCLKWKS234 USCORCRFWKS004 USCORCRFWKS040 USCORCRFWKS069 USCORCRFWKS070 USCORCRFWKS074 Current Frontend Release 3.50 3.50 3.50 3.50 3.50 3.50 Current Frontend Patch 6 6 6 6 6 6 Recommended Frontend Release Recommended Frontend Patch Rating 3.50 3.50 3.50 3.50 3.50 3.50 9 9 9 9 9 9 Red Red Red Red Red Red 12.7.2 PCs with old SAP GUI installation Available SAP GUI releases for your BW release 6.20 6.40 EarlyWatch Alert Report , , 8.03.2007 SAP GUI version support timeline SAP GUI 6.20 support is planned to be "long term". SAP GUI 6.40 support ends on December 31, 2007. 37 Trend Analysis 13 Trend Analysis This section contains the trend analysis for key performance indicators (KPIs). Diagrams or tables are built week by week after activating the EarlyWatch Alert Service. 13.1 System Activity The following diagrams or tables contain data about system activity: - Total Activity: Transaction steps performed each week (in thousands) - Dialog Activity: Transaction steps performed in dialog task each week (in thousands) - Peak Activity: Transaction steps in hour during the peak hour; this peak hour is calculated as the hour with the maximum dialog activity in the ST03 time profile divided by 5 (work days) - Total Users: Users that logged on in the week - Active Users: Users that have performed more than 400 transaction steps in the week 13.2 Response Times The following diagram or table show the development of the response times: - Average response time in dialog task (total, DB and CPU parts) - Average response times for database requests - Average response times for the top 5 transactions EarlyWatch Alert Report , , 8.03.2007 38 Trend Analysis Transaction Code RSA1 SESSION_MANAGER RSPCM RSANWB RSAN_WB_TST Load (%) 38.1 15.7 13.2 10.7 8.0 13.3 System Operation The following diagram or table shows important KPIs for system operation. Note: The column 'Availability' is filled for systems with SAP Basis 4.6 and higher. EarlyWatch Alert Report , , 8.03.2007 39 Trend Analysis 13.4 Hardware Capacity EarlyWatch Alert Report , , 8.03.2007 40