From nobody@FreeBSD.org  Tue Feb 13 06:59:40 2001
Return-Path: <nobody@FreeBSD.org>
Received: from freefall.freebsd.org (freefall.freebsd.org [216.136.204.21])
	by hub.freebsd.org (Postfix) with ESMTP id 6B7C137B65D
	for <freebsd-gnats-submit@FreeBSD.org>; Tue, 13 Feb 2001 06:59:40 -0800 (PST)
Received: (from nobody@localhost)
	by freefall.freebsd.org (8.11.1/8.11.1) id f1DExeM17860;
	Tue, 13 Feb 2001 06:59:40 -0800 (PST)
	(envelope-from nobody)
Message-Id: <200102131459.f1DExeM17860@freefall.freebsd.org>
Date: Tue, 13 Feb 2001 06:59:40 -0800 (PST)
From: elian@xs4all.nl
To: freebsd-gnats-submit@FreeBSD.org
Subject: uid0 processes get killed after getswapspace failed caused by normal user process
X-Send-Pr-Version: www-1.0

>Number:         25062
>Category:       kern
>Synopsis:       uid0 processes get killed after getswapspace failed caused by normal user process
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Tue Feb 13 07:00:01 PST 2001
>Closed-Date:    Sat Jun 2 02:43:10 PDT 2001
>Last-Modified:  Sat Jun 02 02:43:33 PDT 2001
>Originator:     Elian de Jager
>Release:        FreeBSD 4.2-RELEASE #0
>Organization:
Recess Metro
>Environment:
FreeBSD recess.metro 4.2-RELEASE FreeBSD 4.2-RELEASE #0: Tue Jan  2 12:22:00 CET 2001     eli@recess.metro:/usr/src/sys/compile/METRO  i386
>Description:
Several processes were killed apparently by running out of swapspace. The application used was "convert" from the "ImageMagick" package which we use for tiff>pdf translation. This process ran under normal uid. 

A summery from systctl -a:
[...]
machdep.msgbuf: swap_pager_getswapspace: failed
swap_pager_getswapspace: failed  (repeated several times}
<3>pid 67939 (john), uid 1003, was killed: out of swap space
swap_pager_getswapspace: failed
<3>pid 69085 (top), uid 0, was killed: out of swap space
swap_pager_getswapspace: failed
<3>pid 173 (libhttpd.ep), uid 0, was killed: out of swap space
swap_pager_getswapspace: failed
<3>pid 93 (syslogd), uid 0, was killed: out of swap space
swap_pager_getswapspace: failed
<3>pid 75699 (convert), uid 1003, was killed: out of swap space
swap_pager_getswapspace: failed (repeated several times}
swap_pager_getswapspace: failed
<3>pid 75766 (top), uid 0, was killed: out of swap space
swap_pager_getswapspace: failed
<3>pid 44146 (dnetc), uid 1005, was killed: out of swap space
swap_pager_getswapspace: failed
<3>pid 75771 (cron), uid 0, was killed: out of swap space
<3>pid 75771 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75769 (cron), uid 0, was killed: out of swap space
<3>pid 75769 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75770 (cron), uid 0, was killed: out of swap space
<3>pid 75771 (cron), uid 0, was killed: out of swap space
<3>pid 75765 (convert), uid 1003, was killed: out of swap space
<3>pid 75771 (cron), uid 0, was killed: out of swap space
swap_pager_getswapspace: failed
swap_pager_getswapspace: failed [...]


I dont know if this is normal behaviour under *BSD but I think it would be nicer to kill just that particular memory absorbing process instead of syslogd or any other process not run by that user [etc.]


>How-To-Repeat:
run the batch convert again


>Fix:


>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->closed 
State-Changed-By: kris 
State-Changed-When: Sat Jun 2 02:43:10 PDT 2001 
State-Changed-Why:  
This question comes up regularly on the mailing lists; please 
see the archives for more information. 

http://www.FreeBSD.org/cgi/query-pr.cgi?pr=25062 
>Unformatted:
