From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: linux-nfs-owner@vger.kernel.org Received: from fieldses.org ([174.143.236.118]:41751 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751914Ab1JQNjC (ORCPT ); Mon, 17 Oct 2011 09:39:02 -0400 Date: Mon, 17 Oct 2011 09:39:01 -0400 To: Harald Dunkel Cc: linux-nfs@vger.kernel.org Subject: Re: nfs3 30% faster than nfs4? Message-ID: <20111017133901.GB9809@fieldses.org> References: <4E9BDF01.6090106@aixigo.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <4E9BDF01.6090106@aixigo.de> From: "J. Bruce Fields" Sender: linux-nfs-owner@vger.kernel.org List-ID: On Mon, Oct 17, 2011 at 09:53:37AM +0200, Harald Dunkel wrote: > Hi folks, > > I did a brief test on a local nfs server: Extracting the > linux kernel sources on a nfs partition it seems that nfs3 > is 30% faster than nfs4. For nfs3 I got 5.2MB/sec. Nfs4 > gave me just 3.8MB/sec. The server was idle except for this > test. > > Client and server were running 2.6.39 from the Debian backports > repository. I just changed the mount option on the client to > select nfs3 or nfs4. > > I don't want to complain, but I did not expect this huge loss > in performance for moving to nfs4. Any helpful comment would be > highly appreciated. It might be worth looking at /proc/self/mountstats and seeing if you can figure out where the time's going in the v4 case. (Probably open/close?) That workload is unlikely to be bandwidth-limited so total time is probably the better number to look at instead of bandwidth. --b.