RCA - E2E – Workload Monitor Problems

[Lluis]

Hello forum, after getting again issues with the E2E Workload Monitor i decide to share that info to the forum.

A common issue on workload performance is system available resources, the perfect configuration for table E2E_RESOURCES & Special RFC resources are:

Offical documentation

SOLMAN_DIALOG_WORKPROCESSES:          

Maximum number of work processes on Solution Manager, which can be used by the

Extractor Framework.

SOLMANDIAG:                                        

Used for RFC calls to Java-Stack of Solution  Manager via JCO.

NONE:                                                     

Used for local RFC calls.

Some words from SAP Support friend.

Entries like SM_<SID>00000CLNT<CLNT>_TRUSTED are customer are related to RFC destinations to ABAP systems. Default RESOURCECAP is 1. This is to avoid high load in the source system. I would not recommend to use a value higher than 2.

There are some standard entries. These standard entries are:

RFCRESOURCE                   RESOURCECAP

SOLMAN_DIALOG_WORKPROCESSES:    Value MUST be less than the number of Available dialog processes of the Solution Manager. I would recommend                                                                       minus 4 (in SMP 14 - 4 = 10)

SOLMANDIAG                 2 or 3, maximum 5-> number of WEBADMIN server count MUST be higher. May be the SOLMANDIAG value plus 5, avoid higher values.

NONE                             5 would be fine

<CTS>                           2

Overall this table restricts the usage of resources (dialog processes)

Take care of that table with periocy to to avoid an overload in Solution Manager, that table content change after each managed system setup

How to compress BI data

Normally the problems on E2E workload analysis come from object 0SMD_PE2G as you can see after execution of report SAP_INFOCUBE_DESIGNS.

0SMD_PE2H          /BI0/F0SMD_PE2H     rows: 16.373.372    ratio:

0SMD_PE2H          /BI0/E0SMD_PE2H     rows:          0 ratio:

0SMD_PE2H          /BI0/D0SMD_PE2HP    rows: 544

The bad table has 16 million records ¡, and the /HIO/D0SMD_PE2HP information 544 means that there are 543 partitions, you have to make a compression after that situation and you will work with E2E Workload much faster

For compress data on BW follow that steps:

Transaction RSA1 -> InfoSources -> search (0SMD_PE2H)-> manage -> Compress tab -> Release and wait after job finish correctly

More information about BI compression, from EDW homepage in scn:  http://scn.sap.com/docs/DOC-11982

Regards,

Luis