From owner-freebsd-isp Tue Nov 27 21:15:35 2001 Delivered-To: freebsd-isp@freebsd.org Received: from moat.teksupport.net.au (moat.teksupport.net.au [203.17.1.98]) by hub.freebsd.org (Postfix) with ESMTP id 4B65C37B41A; Tue, 27 Nov 2001 21:15:24 -0800 (PST) Received: from rob.secombe.teksupport.net.au (rob.secombe.teksupport.net.au [192.168.1.2]) by moat.teksupport.net.au (8.11.0/8.11.0) with SMTP id fAS5EnZ98307; Wed, 28 Nov 2001 16:14:49 +1100 (EST) (envelope-from robseco@teksupport.net.au) Message-Id: <3.0.5.32.20011128151448.038174f0@mail.secombe> X-Sender: robseco@mail.secombe X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.5 (32) Date: Wed, 28 Nov 2001 15:14:48 +1000 To: Rowan Crowe , freebsd-isp@FreeBSD.ORG From: Rob Secombe Subject: Re: Whats with this -> sendto: No buffer space available Cc: freebsd-stable@FreeBSD.ORG In-Reply-To: References: <3C0454B6.6000102@quake.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-isp@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hi, I agree. I believe it is as a result of the output queue filling as a result of the link dying. In my case our customer was using a Alcatel ADSL modem and it was the ethernet link between the FreeBSD box and the modem that was falling over. I suspect that the NIC was trying to negotiate 100baseTX or the like and the modem is only capable of 10baseT/UTP. The link led on the modem goes out and the media status on the NIC was inactive in the failed state. The problem was resolved thus: ifconfig_rl1="media 10baseT/UTP -mediaopt full-duplex up" I have a bunch of these "out there" running on different providers, including Tel$tra, without any problems. Cheers Rob. At 15:51 28/11/01 +1100, Rowan Crowe wrote: >bOn Wed, 28 Nov 2001, Kal Torak wrote: > >> This keeps happening... >> >> I have adsl running on my gateway using pppoe with userppp >> and its internal nat enabled, also a fairly simple ipfw config >> blocking a few ports... >> >> Now every so often all data stops and checking on the gateway >> trying to send something I get the "sendto: No buffer space available" >> error and have to reboot and everything is fine again... >> >> It seems to happen at random, sometimes its fine for a week, >> sometimes only a few days and it happens... There is nothing >> of interest logged in ppp.log or security or messages... >> >> What causes this buffer to fill up and not empty? Here in australia >> telstras adsl is really crap and you often cant get more than 2 >> hops past the peer... Im wondering if its something to do with >> apps like bitchx trying to constantly reconnect when the network >> is down filling up the buffer??? > >I see this error occasionally on a 56k modem connection. Sometimes the >modem half dies (CD light on but not passing data), and usually drops out >within a few minutes. If there's a lot of outbound data in this period, >the error pops up. > >Presumably the error refers to an output buffer being full, in my case >most likely because the modem has signalled it cannot take any more data, >and ultimately packets cannot be shed from the packet queue. I haven't >used PPPoE but I would guess it means that your client cannot reach the >other end of the tunnel, which is effectively the same thing - output >queue fills and overflows. > >Cheers. > > >-- >Rowan Crowe >camrecord.com / camdiscover.com / Sensation Internet Services >Melbourne, Australia > > > > > >To Unsubscribe: send mail to majordomo@FreeBSD.org >with "unsubscribe freebsd-isp" in the body of the message > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message