OT: high WIO on Linux

Sorry if this is a bit off topic, but I imagine that a few of you may have encountered this in the past...

Recently, I have observed a couple different Linux boxes (I'm reasonably sure both were RHEL3) displaying very unusual-looking WIO statistics. On one, I was seeing about 15% USER, and 80%+ WIO. (Okay, that's bad, but not preposterous, except that 'iostat' showed all disks were essentially idle!) The other was even weirder -- 50% IDLE and 50% WIO. There was nothing there to generate I/O requests.

Both machines had (at least) one thing in common -- aside from running Oracle. They each had *much* more RAM than SWAP. (RAM/SWAP was 6GB/2GB for the first case, and 4GB/1GB for the second.)

Is it possible that all of these I/O waits are being generated inside the kernel -- perhaps by the filesystem buffer cache trying to steal pages from itself, or something equally unproductive.

By the way, in neither case was there noticeable evidence of paging or swapping. The swap space itself was esentially unused...

Can anybody offer a plausible (even if it is hypothetical) explanation for this behaviour? Or maybe direct me to some resources that will help me better understand *how* or *why* such things can happen?

I mean, really! 50% IDLE + 50% WIO is awfully strange! My Linux skills are maybe a bit "lightweight", but I've been working with UNIX for decades and never seen (or at least never *noticed*) anything remotely like this...

Cheers,
-- Mark.

--

http://www.freelists.org/webpage/oracle-l Received on Fri Mar 10 2006 - 01:05:11 CST