First Twitter DB2 Q&A Session – Top Questions!

Thanks to all our Twitter followers who joined us for our first DB2 Geek Q&A Session.  We had some great questions which got our Geeks thinking!  Below are a selection of our top questions and answers.  Our next Q&A session is scheduled for 21st May 4pm BST.  If you can’t attend on the day, send your question in early, we will feature it during the live session and send your answer direct to you.  Follow #db2geek on Twitter to see all of the questions and answers.

Q. Any idea if DB2 z/OS V11 will provide the same levels of CPU savings as V10?

A. It depends 🙂 DB2 10 had some great CPU savings for many customers, and will be a tough act to follow.  DB2 11 contains additional features to benefit query workloads, heavy INSERT, SAP, read of compressed tables and others.  It is still too early to tell what “typical” CPU savings will be but we would expect them to be somewhat less than DB2 10 for most but still better than the 5-10% *regression* typical of previous upgrades.

Q. Why would DB2 be obsessed with running run stats multiple times a day on a table, data only changes once a day AUTO_RUNSTATS = ON?

A. You can set a maintenance window for auto runstats as part of the profile, so if the data changes once a day limit the window.  However the UDI value used to kick off autostats shouldn’t change if table isn’t updated, so you shouldn’t see this behaviour

Q. Do you have any strategies for dealing with long running SQL statements?

A. 1st rule of long running SQL statements is to check the access path via Explain. Also ensure catalog stats are up to date.  Put the SQL through DB2 explain to see what access path is being used and db2advis to find potential  improvements.  Ensure no missing joins, and try to reduce number of qualifying rows as early as possible.

Follow us @db2geek

• Share