From owner-freebsd-sparc64@FreeBSD.ORG Fri Oct 1 13:38:54 2004 Return-Path: Delivered-To: freebsd-sparc64@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C99D016A4CE for ; Fri, 1 Oct 2004 13:38:54 +0000 (GMT) Received: from catflap.slightlystrange.org (cpc2-cmbg1-3-0-cust94.cmbg.cable.ntl.com [213.107.104.94]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5A86E43D5C for ; Fri, 1 Oct 2004 13:38:54 +0000 (GMT) (envelope-from www@slightlystrange.org) Received: from www by catflap.slightlystrange.org with local (Exim 4.42 (FreeBSD)) id 1CDNca-0000O9-M2 for freebsd-sparc64@freebsd.org; Fri, 01 Oct 2004 14:38:52 +0100 Received: from 192.168.0.1 (proxying for 127.0.0.1) (SquirrelMail authenticated user dan); by catflap.slightlystrange.org with HTTP; Fri, 1 Oct 2004 14:38:52 +0100 (BST) Message-ID: <49894.192.168.0.1.1096637932.squirrel@192.168.0.1> Date: Fri, 1 Oct 2004 14:38:52 +0100 (BST) From: "Daniel Bye" To: freebsd-sparc64@freebsd.org User-Agent: SquirrelMail/1.4.3a X-Mailer: SquirrelMail/1.4.3a MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-15 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal Sender: World Wide Web Owner Subject: Runaway bash on FreeBSD SPARC64 X-BeenThere: freebsd-sparc64@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: freebsd@slightlystrange.org List-Id: Porting FreeBSD to the Sparc List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Oct 2004 13:38:54 -0000 Hi folks, Not entirely sure this is the correct forum - if not, then a gentle nudge in the right direction would be appreciated! I've searched the list archives and had a scan of Google, but nothing appropriate came back. I am running FreeBSD/SPARC64 5.2.1-RELEASE-p10 on an Ultra 10. On occasions, bash(1) goes insane and grabs ever more memory. Other processes get forced out to swap, and bash itself eventually occupies all physical RAM and starts paging as well. Most recently, I noticed bash's SIZE in top reach 1031MB - which is well on the way to being all available physical RAM and swap space. Once it gets like this, bash fails to respond to kill(1), and the only way to get normal service back is to force a reboot. All I could find in /var/log/messages after having to reboot this morning was this: Oct 1 11:30:58 catflap shutdown: reboot by danielby: Oct 1 11:31:16 catflap kernel: swap_pager: out of swap space Which really doesn't say a lot. There was a bash.core file dropped in the shell's cwd, but got removed by an over-zealous cleanup script. There appears to be no pattern involved - it can happen at any time, no matter which user runs it (a real user or root), or what they are doing in the shell at the time. I run FreeBSD on i386 and AMD64 as well, and never see this behaviour. bash(1) is installed from the ports - currently version bash-2.05b.007_2 Anyone else seen this? Or have any hints or suggestions on what I can do to figure where the problem is? I'm no hacker, but am willing to put in some effort if it's required. Thanks all, Dan -- Daniel Bye PGP Key: ftp://ftp.slightlystrange.org/pgpkey/dan.asc PGP Key fingerprint: 3B9D 8BBB EB03 BA83 5DB4 3B88 86FC F03A 90A1 BE8F _ ASCII ribbon campaign ( ) - against HTML, vCards and X - proprietary attachments in e-mail / \