From owner-freebsd-net Wed Apr 28 14: 7:50 1999 Delivered-To: freebsd-net@freebsd.org Received: from ss1000.ms.mff.cuni.cz (ss1000.ms.mff.cuni.cz [195.113.19.221]) by hub.freebsd.org (Postfix) with ESMTP id 4C42614F7B for ; Wed, 28 Apr 1999 14:07:47 -0700 (PDT) (envelope-from mkop5230@ss1000.ms.mff.cuni.cz) Received: from beta.ms.mff.cuni.cz (mkop5230@beta.ms.mff.cuni.cz [195.113.16.70]) by ss1000.ms.mff.cuni.cz (8.8.8/8.8.8) with ESMTP id XAA08138 for ; Wed, 28 Apr 1999 23:07:47 +0200 Received: from localhost (mkop5230@localhost) by beta.ms.mff.cuni.cz (980427.SGI.8.8.8/8.8.8) with ESMTP id XAA11597 for ; Wed, 28 Apr 1999 23:07:46 +0200 (MDT) Date: Wed, 28 Apr 1999 23:07:46 +0200 From: Milan Kopacka Reply-To: Milan Kopacka To: freebsd-net@freebsd.org Subject: Out of mbuf clusters Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hallo, I'm getting "Out of mbuf clusters" kernel panic, if there are processes in the system dying and leaving many opened tcp sockets. Special attack -> forking "open tcp socket and die" processes needs time under 30 sec to bring the machine down (thousands of opened sockets). FreeBSD 3.0 or 3.1; no special tuning of GENERIC kernel, just disabling some hw drivers & so on. (K6/300, 64MB of memory, de0 eth driver) In mailing lists archives I have found an advice to tune the kernel by increasing NMBCLUSTERS. How much does it help to raise the flag higher? I'm not talking about normal load, but about such mbuf attack from non-root user. Could someone explain to me, why being out of mbuf clusters is a reason to kernel panic? thanks Milan Kopacka -- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message