From owner-freebsd-current@FreeBSD.ORG Mon Nov 24 15:43:08 2003 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 6459716A4CE; Mon, 24 Nov 2003 15:43:08 -0800 (PST) Received: from kai.xtaz.net (82-32-25-111.cable.ubr04.azte.blueyonder.co.uk [82.32.25.111]) by mx1.FreeBSD.org (Postfix) with ESMTP id 624BE43FBF; Mon, 24 Nov 2003 15:43:07 -0800 (PST) (envelope-from matt@xtaz.net) Received: from xtaz.net (xai.xtaz.net [10.0.0.2]) by kai.xtaz.net (Postfix) with ESMTP id 628EB90340; Mon, 24 Nov 2003 23:43:06 +0000 (GMT) Message-ID: <3FC2978A.4020901@xtaz.net> Date: Mon, 24 Nov 2003 23:43:06 +0000 From: Matt Smith User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.5) Gecko/20031101 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Sam Leffler References: <3FC1F7E6.40608@xtaz.net> <200311240930.14810.sam@errno.com> In-Reply-To: <200311240930.14810.sam@errno.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: Robert Watson cc: current@freebsd.org cc: sos@deepcore.dk Subject: Re: NFS lockup issues and xl0 watchdog timeout X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 24 Nov 2003 23:43:08 -0000 I've had a possible idea regarding the NFS issues. I'm wondering if perhaps my NFS issues are related to the other email thread I have going which is the xl0: watchdog timeouts etc. I had not noticed this until last week because it's not often I copy large files from one machine to another but doing an ftp/scp etc from one machine to the other I'm only getting 100KB/sec with a PUT from my nfsclient->nfsserver, and 6MB/sec with a GET. This used to go at 9-10MB/sec both ways. Thinking about it I think this started happening around the same time I started having NFS issues. My NFS issues are also only when writing to the server from the client, never the other way around. So this ties up with the throughput figure difference I get on FTP. Jimmy mentioned a watchdog timeout caused by what looked like a 10-15 second delay in a buffer going from userland to kernel? I guess if this is the case it could kill NFS as well. Tomorrow night I will change my xl0 card in this desktop for a spare realtek card I happen to have and test throughput and NFS. I'll let you know my results. It would be interesting to see if the other person who reported these NFS issues (soren) has a similar experience. Matt.