From owner-freebsd-net@FreeBSD.ORG Wed Mar 30 17:59:58 2011 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 84FBE106566B for ; Wed, 30 Mar 2011 17:59:58 +0000 (UTC) (envelope-from andrey@zonov.org) Received: from mail-fx0-f54.google.com (mail-fx0-f54.google.com [209.85.161.54]) by mx1.freebsd.org (Postfix) with ESMTP id 1CC0A8FC0C for ; Wed, 30 Mar 2011 17:59:57 +0000 (UTC) Received: by fxm11 with SMTP id 11so1610673fxm.13 for ; Wed, 30 Mar 2011 10:59:57 -0700 (PDT) Received: by 10.223.101.87 with SMTP id b23mr1668324fao.97.1301507996917; Wed, 30 Mar 2011 10:59:56 -0700 (PDT) Received: from [10.254.254.77] (ppp95-165-144-57.pppoe.spdop.ru [95.165.144.57]) by mx.google.com with ESMTPS id 17sm129402far.43.2011.03.30.10.59.54 (version=SSLv3 cipher=OTHER); Wed, 30 Mar 2011 10:59:55 -0700 (PDT) Message-ID: <4D936F99.3060508@zonov.org> Date: Wed, 30 Mar 2011 21:59:53 +0400 From: Andrey Zonov User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ru; rv:1.8.1.24) Gecko/20100228 Thunderbird/2.0.0.24 Mnenhy/0.7.6.0 MIME-Version: 1.0 To: Arnaud Lacombe References: <4D923931.2070606@zonov.org> <4D92C673.2080107@zonov.org> In-Reply-To: Content-Type: text/plain; charset=KOI8-R; format=flowed Content-Transfer-Encoding: 8bit Cc: freebsd-net@freebsd.org, Jack Vogel Subject: Re: igb(4) won't start with "igb0: Could not setup receive structures" X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 30 Mar 2011 17:59:58 -0000 Hi, Maybe you're right. OK, let's return default hw.igb.rxd to 256. It seems to be enough for stable work and driver is used less memory. BTW, on the man page igb(4) still written that hw.igb.rxd equals to 256 by default. -- Andrey Zonov 30.03.2011 18:33, Arnaud Lacombe пишет: > Hi, > > On Wed, Mar 30, 2011 at 1:58 AM, Andrey Zonov wrote: >> My point is if you're using machine with 8 CPUs than maxusers/clusters/9k >> mbufs should have been increased by system, because on this machine minimum >> 2Gb memory is available. >> > I am doubtful that the number of CPU[0] or number of users (yes, I > know `maxusers' is currently used to compute the default > `nmbcluster'...) can be linked to any network load pattern at all. You > can have a 24 CPU machine made for 4096 users with a single NIC, not > requiring much memory, while a 1 CPU machine with only 1 users can > have +8 NIC and require a huge quantity of memory. Available KVM space > should also be taken into account, as it is rather limited on i386. > > - Arnaud > > [0]: even more today where you can have a huge number of virtual CPU.