From nobody@FreeBSD.ORG Fri Mar 19 13:49:46 1999
Return-Path: <nobody@FreeBSD.ORG>
Received: by hub.freebsd.org (Postfix, from userid 32767)
	id C9D491523B; Fri, 19 Mar 1999 13:49:45 -0800 (PST)
Message-Id: <19990319214945.C9D491523B@hub.freebsd.org>
Date: Fri, 19 Mar 1999 13:49:45 -0800 (PST)
From: ath@niksun.com
Sender: nobody@FreeBSD.ORG
To: freebsd-gnats-submit@freebsd.org
Subject: 3.1-RELEASE deadlocks under load with processes in vmwait
X-Send-Pr-Version: www-1.0

>Number:         10676
>Category:       kern
>Synopsis:       3.1-RELEASE deadlocks under load with processes in vmwait
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Mar 19 13:50:01 PST 1999
>Closed-Date:    Thu Nov 4 10:06:17 PST 1999
>Last-Modified:  Thu Nov  4 10:06:56 PST 1999
>Originator:     Andrew Heybey
>Release:        3.1-RELEASE
>Organization:
Niksun, Inc.
>Environment:
FreeBSD alpha.niksun.com 3.1-RELEASE FreeBSD 3.1-RELEASE #1: Wed Mar 17 15:06:52 EST 1999     ath@alpha.niksun.com:/usr/src/sys/compile/ALPHA  i386
>Description:
Under heavy network and disk load, the machine locks up (or at
least any process that wants some pages gets stuck in vmwait).

ddb reveals many processes in "vmwait" with others (including
the pagedaemon() in "inode" wait.
>How-To-Repeat:
I have two fxp fastethernets each receiving about 15k 512-byte pkts/sec.
The above data (~15MB/sec) is all being written to a ccd partition
striped across three disks.
Several processes are also reading the above data.

>Fix:


>Release-Note:
>Audit-Trail:

From: Andrew Heybey <ath@niksun.com>
To: freebsd-gnats-submit@freebsd.org
Cc:  
Subject: kern/10676: 3.1-RELEASE deadlocks under load with processes in vmwait
Date: Thu, 04 Nov 1999 11:30:22 -0500

 I just noticed this was still open, so I thought that I would add that 
 this problem was fixed by commits made to the vm system right about
 the date of the PR by Julian Elischer.  Sorry, I don't remember the exact
 changes but I know that by upgrading my world to -STABLE as of 21-MAR-99 
 the problem went away.
 
State-Changed-From-To: open->closed 
State-Changed-By: ken 
State-Changed-When: Thu Nov 4 10:06:17 PST 1999 
State-Changed-Why:  
Submitter reports that his problem has gone away. 
>Unformatted:
