I am commonly asked why I don't look at the pages swapped in and only the pages swapped out. Well, if there are pages being swapped in, there must have been pages swapped out. And if pages have been swapped out, there is a good chance they will eventually be paged back in.

Figure 4-15. A vmstat report on a very active four-CPU core database server. The so column represents the number of pages swapped out per second during the reporting interval. As in this sample, we want to this nearly always to be zero.

Performance analysts know that if an Oracle server process (and, God forbid, a background process) has some of its memory pages swapped out, response time for that process will be negatively impacted if the pages are swapped back in. And let's not forget the pages were swapped out because another active process requested memory and the operating system could not immediately provide the memory.