From owner-freebsd-questions@FreeBSD.ORG Fri Jan 18 21:10:24 2013 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by hub.freebsd.org (Postfix) with ESMTP id 2624E20F for ; Fri, 18 Jan 2013 21:10:24 +0000 (UTC) (envelope-from realrichardsharpe@gmail.com) Received: from mail-vb0-f47.google.com (mail-vb0-f47.google.com [209.85.212.47]) by mx1.freebsd.org (Postfix) with ESMTP id D30B381D for ; Fri, 18 Jan 2013 21:10:23 +0000 (UTC) Received: by mail-vb0-f47.google.com with SMTP id e21so4036334vbm.6 for ; Fri, 18 Jan 2013 13:10:17 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:x-received:date:message-id:subject:from:to :content-type:content-transfer-encoding; bh=fU1NMEA9Zssq0STfdQ+IheM/3+wt5KJHVMedtVFAarU=; b=Iw9dRyW3aAdOblr1cgD94aC4OWrarZJsx2zneurHGr0yxubfcJc4xbDVUmAW8CE6Y1 aUoRSj9Nj2xFSFyhyNdOOQgD4jJTn6ado3vAjiHwpQFcGj87XetviQ2yz1PUmq9iHBGg mkXBp7+VciK12U6Ln0OOBwE5JJnmzuauhNm2Lyiy5ZajG4agobnApKIBfy9M/7QNbuO6 ELmgUGChO9/+kWp7Cjoq+c9gS3qt3nk/HTEQL3S+9w6XMFFiPOfVFJr4FlVsqUGfsVTz TCXQOlNqyFB2tdok+l2vnrHPkngX3oFP3Sn6BkKDmzUsME7HsTKk+1URdBL6HsnauPqs 4qgA== MIME-Version: 1.0 X-Received: by 10.52.28.176 with SMTP id c16mr9708600vdh.126.1358543417618; Fri, 18 Jan 2013 13:10:17 -0800 (PST) Received: by 10.58.188.78 with HTTP; Fri, 18 Jan 2013 13:10:17 -0800 (PST) Date: Fri, 18 Jan 2013 13:10:17 -0800 Message-ID: Subject: Have there been any fixes for long TCP delays with FreeBSD 8.3 or 9.x? From: Richard Sharpe To: freebsd-questions@freebsd.org Content-Type: text/plain; charset=Big5 Content-Transfer-Encoding: quoted-printable X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Jan 2013 21:10:24 -0000 Hi folks, I am seeing a problem when copying large files via SMB/Samba from a FreeBSD 8.0-based system (with Samba 3.6.6 and ZFS etc) where eventually Windows drops the connection. However, it seems, based on three captures I have, that what has happened is that FreeBSD has not supplied any data on the connection for 50-60 seconds and the Windows 60-second timeout trips. In each case, what I see is a large number of 1500-byte frames go out. Then Windows ACKs a reasonable amount of data, 30+K or so, and then sends another ACK to ack everything so far received. Then in two cases, after about 52 seconds I see some more data in two of the captures I have but after a further few seconds, Windows disconnects. In the last capture, there was no further data for 60 seconds, so Windows disconnected. I know there should have been more data because: 1. The READ requests are for 64kiB and there is still data to come and the sequence numbers match up. 2. In my Samba traces I know that Samba has gone on to process the next SMB2 command in sequence and the time difference between these are of the order of 1mS and Samba in this case is operating in sync mode where it does not read the next command off the socket until it has finished writing the response for the previous command to the socket. The system is relatively idle. Only one smbd and some web traffic (inbound as a result of a get). I am hoping that this describes a known problem. I can provide captures on request. They will have to be cut-down as they are 300 to 500MB in length. --=20 Regards, Richard Sharpe (=A6=F3=A5H=B8=D1=BC~=A1H=B0=DF=A6=B3=A7=F9=B1d=A1C--=B1=E4=BE=DE)