From owner-freebsd-questions@FreeBSD.ORG Tue Oct 5 08:41:09 2004 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 3E77516A4CE for ; Tue, 5 Oct 2004 08:41:09 +0000 (GMT) Received: from rosebud.otenet.gr (rosebud.otenet.gr [195.170.0.26]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8E55643D1F for ; Tue, 5 Oct 2004 08:41:08 +0000 (GMT) (envelope-from keramida@ceid.upatras.gr) Received: from orion.daedalusnetworks.priv (host5.bedc.ondsl.gr [62.103.39.229])i958f6CY024207; Tue, 5 Oct 2004 11:41:06 +0300 Received: from orion.daedalusnetworks.priv (orion [127.0.0.1]) i958f6gT001990; Tue, 5 Oct 2004 11:41:06 +0300 (EEST) (envelope-from keramida@ceid.upatras.gr) Received: (from keramida@localhost)i958f5x4001989; Tue, 5 Oct 2004 11:41:05 +0300 (EEST) (envelope-from keramida@ceid.upatras.gr) Date: Tue, 5 Oct 2004 11:41:05 +0300 From: Giorgos Keramidas To: Joe Schmoe Message-ID: <20041005084105.GA1837@orion.daedalusnetworks.priv> References: <20041004131500.9587.qmail@web53308.mail.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20041004131500.9587.qmail@web53308.mail.yahoo.com> cc: freebsd-questions@freebsd.org Subject: Re: tracking per process IO usage / stats ... help needed.. X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 05 Oct 2004 08:41:09 -0000 On 2004-10-04 06:15, Joe Schmoe wrote: > > I have a FreeBSD system (4.9) running a fair number of processes in a > multi-user / shell hosting environment. > > One problem that routinely comes up is that the system will seem to be > fine in terms of CPU, and none of the top 10 or 15 processes that I > see in "top" are using much CPU ... but the load average is very high > - sometimes as high as 30 or 40 ... > > I have good reason to suspect that the load is high due to I/O. I can > see the number of processes blocking on I/O in vmstat, and I can see > iostat for the entire machine, of course ... You might want to see if the "io" mode of top can be ported to RELENG_4. In 5.X you can use top to display io statistics too: : $ top -m io | sed -e 1,7d | head -10 : PID USERNAME VCSW IVCSW READ WRITE FAULT TOTAL PERCENT COMMAND : 1742 keramida 32250 2182 278 7 1376 1661 28.90% Xorg : 1829 keramida 5617 1601 440 4 472 916 15.94% mozilla-bin : 416 root 5031 5119 0 0 1 1 0.02% moused : 1745 keramida 19871 517 65 0 22 87 1.51% wmaker : 1789 keramida 5264 425 2 0 1 3 0.05% xterm-static : 1837 keramida 643 205 17 188 6 211 3.67% mutt : 1792 keramida 2640 1023 0 0 0 0 0.00% screen : 1949 keramida 260 118 0 16 1 17 0.30% emacs : 293 root 702 181 9 3 0 12 0.21% syslogd The differences of top from RELENG_4 to CURRENT are many and I don't have the time right now to try porting the "io mode" stuff to RELENG_4, but if you cannot do it yourself I might find a bit of time this weekend. - Giorgos