* Jim C. Nasby:
> The problem is that it seems like there's never enough ability to clue
> the OS in on what the application is trying to accomplish. For a long
> time we didn't have a background writer, because the OS should be able
> to flush things out on it's own before checkpoint. Now there's talk of a
> background reader, because backends keep stalling on waiting on disk IO.
I've recently seen this on one of our test systems -- neither CPU nor
disk I/O were maxed out.
Have you considered using asynchronous I/O? Maybe it results in less
complexity and fewer context switches than a background reader.