07.02.2013 Views

Best Practices for SAP BI using DB2 9 for z/OS - IBM Redbooks

Best Practices for SAP BI using DB2 9 for z/OS - IBM Redbooks

Best Practices for SAP BI using DB2 9 for z/OS - IBM Redbooks

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Use transaction SM21 to check your <strong>SAP</strong> system logs <strong>for</strong> any error situations<br />

that may be ca<strong>using</strong> per<strong>for</strong>mance problems. Use the menu path<br />

Systemlog → Choose → All remote system locks.<br />

2. Check application server health.<br />

If there are no obvious errors on the <strong>SAP</strong> system logs, begin checking the<br />

health of the application servers:<br />

a. Use transaction ST06 to check the CPU consumption. From the menu<br />

path Goto → Current data → Snapshot → Top CPU processes, you<br />

can determine whether a user transaction or report is looping and <strong>using</strong> all<br />

of the available CPU.<br />

b. Check the health of the <strong>SAP</strong> buffers <strong>using</strong> transaction ST02.<br />

3. Check the database connection.<br />

If there are no problems with the servers in question, confirm that the<br />

database connection from the server to the <strong>DB2</strong> data sharing member is<br />

per<strong>for</strong>ming well:<br />

a. Check the DRDA® network connection by <strong>using</strong> <strong>SAP</strong> transaction <strong>DB2</strong>,<br />

selecting the <strong>DB2</strong> Per<strong>for</strong>mance Tuning tab, then selecting <strong>DB2</strong> ping to<br />

test the network response time.<br />

This should show a response time of 1–2 milliseconds to send 500 bytes<br />

of data. If this response time is longer, then there is a problem with the<br />

link.<br />

You can investigate this further by clicking <strong>DB2</strong> connect diagnostics on<br />

the <strong>DB2</strong> Per<strong>for</strong>mance Tuning tab.<br />

b. Continue the investigation by asking the network support group to<br />

investigate the problem.<br />

4. Check the z/<strong>OS</strong> database server health.<br />

Per<strong>for</strong>mance analysis on <strong>OS</strong>/390® z/<strong>OS</strong> is beyond the scope of this book, but<br />

there are some basic checks that you can per<strong>for</strong>m:<br />

– Check the <strong>OS</strong>/390 z/<strong>OS</strong> system logs <strong>for</strong> any error situations that may<br />

cause problems on the LPAR.<br />

– Check the <strong>DB2</strong> MSTR job log <strong>for</strong> any error situations in <strong>DB2</strong> that may be<br />

ca<strong>using</strong> per<strong>for</strong>mance problems.<br />

– Check the LPAR CPU utilization and look <strong>for</strong> looping address spaces<br />

consuming CPU.<br />

– Use RMF Monitor III to check your workloads <strong>for</strong> any delay.<br />

Chapter 8. Query and Load per<strong>for</strong>mance 139

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!