This needs a longer answer than I have time but the short version is 1. just increasing the buffer pools because you have more memory may make more problems than it fixes. There is no point in saving a block in a buffer pool if there isn't another task/transaction coming along to use it. 2. If you haven't defined different DB2 buffer pools by usage, you should do that first. Index buffer pools, data buffer pools, catalog buffer pool, and temp buffer pool is the minimum starting point. After that you should "consider" making bufferpools for batch vs online so the big batch processes don't trash your online transaction response time. 3. Assuming you are on DB2 V8, you should consider page fixing the highest read/write activity buffer pools. The page fix/release is much more expensive for 64 bit addressing. The buffer pools that have the most read/write activity should be page fixed. You will find these counts in the SMF101 stats records. 4. change the default thresholds for prefetch and write activity for the buffer pools. the defaults are awfull. 5. Once you have gotten that far consider something like Joel Goldsteins buffer pool tool. It will analyze your usage patterns and identify which buffer pools need more memory. He can always reduce the IO by 30% and many times more than that.
Mike -- Mike ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
