From owner-freebsd-alpha Sun Feb 17 13:33:26 2002 Delivered-To: freebsd-alpha@freebsd.org Received: from srv1.cosmo-project.de (srv1.cosmo-project.de [213.83.6.106]) by hub.freebsd.org (Postfix) with ESMTP id 3D0D137B402 for ; Sun, 17 Feb 2002 13:33:22 -0800 (PST) Received: (from uucp@localhost) by srv1.cosmo-project.de (8.11.6/8.11.6) with UUCP id g1HLXFm07300; Sun, 17 Feb 2002 22:33:15 +0100 (CET) (envelope-from ticso@cicely8.cicely.de) Received: from cicely8.cicely.de (cicely8.cicely.de [10.1.2.10]) by cicely5.cicely.de (8.12.1/8.12.1) with ESMTP id g1HLWueS021902; Sun, 17 Feb 2002 22:32:56 +0100 (CET)?g (envelope-from ticso@cicely8.cicely.de) Received: (from ticso@localhost) by cicely8.cicely.de (8.11.6/8.11.6) id g1HLWul43736; Sun, 17 Feb 2002 22:32:56 +0100 (CET) (envelope-from ticso) Date: Sun, 17 Feb 2002 22:32:56 +0100 From: Bernd Walter To: Andrew Gallatin Cc: freebsd-alpha@freebsd.org Subject: Re: runningbufspace deadlock problem on -stable Message-ID: <20020217213256.GA43658@cicely8.cicely.de> References: <20020217180949.GD42870@cicely8.cicely.de> <15472.5308.590348.105307@grasshopper.cs.duke.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <15472.5308.590348.105307@grasshopper.cs.duke.edu> User-Agent: Mutt/1.3.26i X-Operating-System: FreeBSD cicely8.cicely.de 5.0-CURRENT i386 Sender: owner-freebsd-alpha@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On Sun, Feb 17, 2002 at 03:38:20PM -0500, Andrew Gallatin wrote: > > Bernd Walter writes: > > I have one -stable production PC164 machine with the problem that > > runningbufspace increases over time. > <..> > > vfs.runningbufspace: 120832 > > > > The value should return to 0 when the box is IO idle. > > > > If someone sees this problems on a local machine I would be happy. > > I have some patches here adding assertions which I want to avoid > > trying on my production machine. > > I recently had a problem on my workstation. (UP1000, 640MB, 4.5-STABLE, > also an NFS server for my various testboxes in my home office). I was > doing a local cvs diff on the src/sys tree, as well as building the > TGA kernel, and doing a few other things (xemacs, gnome, sawfish, > konqueror, 20 or so shells). The cvs diff got wedged in (I think) > inode. State inode sounds to be something different. But the important part is the state of bufdaemon. > At this point, I noticed that the FFS node malloc pool seemed to be > quite near its limit. I killed the make and tried to recover, but I > couldn't seem to get the number of FFS node allocations down & other > jobs started to wedge on IO. I was intending to drop into the > debugger and get a dump, but the machine locked solid when I > tty-switched out of X. If it's the runningbufspace thing a dump is only describing the effect not the reason. > I have so far be unable to reproduce the problem. > > I wonder if these two problems might be related..? With DDB's ps you will see bufdaemon and possibly other processes blocked in state wdrain. You might take a look at vfs.runningbufspace and see if it's greater than 0 when the system is idle. It's increasing on my box several times a day so you might see it quite early. Once it's raised over vfs.lorunningspace (512k default) async writes blocked in wdrain never wake up. My insterest is to trap the situation where runningbufspace handling does something wrong. -- B.Walter COSMO-Project http://www.cosmo-project.de ticso@cicely.de Usergroup info@cosmo-project.de To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message