Re: Determining what is causing high wio%

From: Allen Kirby <akirby_at_att.com>
Date: 1996/12/05
Message-ID: <32A6D4E2.2051_at_att.com>#1/1


Jonathan Lewis wrote:
>
> Allen,
>
> Is it possible that the I/O for which Suns are waiting is actually
> I/O relating to inter process communication, rather than 'real
> disk' I/O ?
>
> ---
> Jonathan Lewis.
> ora_mail_at_jlcomp.demon.co.uk

I don't think so. IPC is done in memory, and cpus don't have to 'wait' on memory like they wait on disks. Processes will wait on IPC resources (like waiting for a semaphore to become > 0) but the cpu itself will process other requests. I don't have access to my Solaris machine right now, but the HP version (9.x) of sar defines wio as raw I/O, block I/O and Virtual Memory pageins/swapins. I would assume Sun is the same or close. Just do 'man sar' and see what the Solaris man page says. Or check the answerbook to be sure.

-- 
---
Allen Kirby			AT&T ITS Production Services
akirby_at_att.com			Alpharetta, GA.
Received on Thu Dec 05 1996 - 00:00:00 CET

Original text of this message