From mboxrd@z Thu Jan 1 00:00:00 1970 From: "J. Bruce Fields" Subject: Re: [PATCH] Exportfs and rpc.mountd optimalization (revisited) Date: Fri, 20 Feb 2009 13:31:46 -0500 Message-ID: <20090220183146.GD1104@fieldses.org> References: <499C5861.6050408@RedHat.com> <20090219174118.GA15743@fieldses.org> <499EC930.2030408@RedHat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Linux NFS Mailing list To: Steve Dickson Return-path: Received: from mail.fieldses.org ([141.211.133.115]:49441 "EHLO pickle.fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753414AbZBTSbq (ORCPT ); Fri, 20 Feb 2009 13:31:46 -0500 In-Reply-To: <499EC930.2030408-AfCzQyP5zfLQT0dZR+AlfA@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Fri, Feb 20, 2009 at 10:16:00AM -0500, Steve Dickson wrote: > > > J. Bruce Fields wrote: > > On Wed, Feb 18, 2009 at 01:50:09PM -0500, Steve Dickson wrote: > >> Very long time ago Tomas Richter submitted a patch that added > >> a export hash table to both exportfs and mountd. Very recently > >> I ran across it and actually used some parts in the recent TCP > >> wrappers patch... > >> > >> Looking at the discussion > >> http://marc.info/?l=linux-nfs&m=111938362106991&w=2 > >> It seem no one took the time to test out the patch.. There > >> was even a Red Hat bug opened on it but for some reason > >> it got closed... Anyway... > >> > >> This is a good patch... imho.. It does help with very large exports > >> at least in my testing... > > > > Just out of curiosity--what's your test? > Created a couple hundred exports. Then used the 'time' command to > see how long 'exportfs -arv' took. Plus I timed how long it took a > client mount... I did to this a while ago, but as I remember it > was a big win with this patch... Sounds good. But I'm surprised a couple hundred is all it takes--that's not such a long list to search! It would be interesting to take a closer look and do some profiling. --b.