From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 6161 invoked by alias); 27 Nov 2008 07:05:21 -0000 Received: (qmail 6141 invoked by uid 22791); 27 Nov 2008 07:05:19 -0000 X-Spam-Check-By: sourceware.org Received: from netic.com (HELO netic.com) (208.131.130.247) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 27 Nov 2008 07:04:35 +0000 Received: from piii550 (60-242-91-36.static.tpgi.com.au [60.242.91.36]) by netic.com (8.13.1/8.13.1) with ESMTP id mAR74UEB006678 for ; Thu, 27 Nov 2008 00:04:32 -0700 Reply-To: From: "Laurie Gellatly" To: References: <333179.53101.qm@web94613.mail.in2.yahoo.com> <30c102240811261217g68b29d4fh43f37e1d15a8c492@mail.gmail.com> <005f01c95009$d87713a0$89653ae0$@gellatly@netic.com> In-Reply-To: <005f01c95009$d87713a0$89653ae0$@gellatly@netic.com> Date: Thu, 27 Nov 2008 13:16:00 -0000 Message-ID: <003c01c9505e$661f3940$325dabc0$@gellatly@netic.com> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Content-Language: en-au X-IsSubscribed: yes Mailing-List: contact ecos-discuss-help@ecos.sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: ecos-discuss-owner@ecos.sourceware.org Subject: RE: [ECOS] Sockets and files X-SW-Source: 2008-11/txt/msg00214.txt.bz2 > -----Original Message----- > From: ecos-discuss-owner@ecos.sourceware.org [mailto:ecos-discuss- > owner@ecos.sourceware.org] On Behalf Of Laurie Gellatly > Sent: Thursday, 27 November 2008 7:59 AM > To: ecos-discuss@ecos.sourceware.org > Subject: [ECOS] Sockets and files > > Hi All, > My eCos project with the FreeBSD stack is undergoing some more real > world tests (put into a more typical network environment). > The network related threads are a web server and an SNMP agent. > Lately I've noticed that the web server thread does not accept > connections for about 1 minute after startup. > It used to accept connections within a few seconds. > To avoid running out of sockets, (which seems to be a recurring problem > I've seen in this forum) > I've had to double CYGPKG_NET_MAXSOCKETS and then increase > CYGPKG_NET_MEM_USAGE to > larger than its default calculation and change CYGPKG_NET_MEMPOOL_SIZE > and CYGPKG_NET_MBUFS_SIZE to their new defaults. > Initially this did not help until I also increased FOPEN_MAX, > CYGNUM_FILEIO_NFILE and CYGNUM_FILEIO_NFD to be almost > the same as the number of sockets. > > I'd like to understand the relationship between these values (sockets > and file opens) so I can set these values > by more than just trial and error and possibly find why the 1 minute > delay occurs. > > Can someone point me at some documentation that explains these > relationships or offer advice? > Just a quick update, the 1 minute delay was being imposed by my switch. At least that's one solved. Thanks ...Laurie:{) -- Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss