From: Ric Wheeler Subject: Re: [PATCH] imporve jbd2 fsync batching Date: Thu, 06 Nov 2008 07:45:38 -0500 Message-ID: <4912E6F2.6010705@redhat.com> References: <20081104161024.GC28058@unused.rdu.redhat.com> <20081105231012.GJ3184@webber.adilger.int> <20081106002727.GB8168@mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Andreas Dilger , Josef Bacik , linux-ext4@vger.kernel.org To: Theodore Tso Return-path: Received: from mx2.redhat.com ([66.187.237.31]:44901 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753450AbYKFMqC (ORCPT ); Thu, 6 Nov 2008 07:46:02 -0500 In-Reply-To: <20081106002727.GB8168@mit.edu> Sender: linux-ext4-owner@vger.kernel.org List-ID: Theodore Tso wrote: > On Wed, Nov 05, 2008 at 04:10:12PM -0700, Andreas Dilger wrote: > >> There was no reply to my previous comments about making the maximum sleep >> time be a fixed value (e.g. 15ms) instead of having it arbitrarily based >> on the jiffies value, which may change between 1ms and 10ms. >> > > I know the whole point is to make it be self-tuning, but I wonder if > we should have some knobs to control the min and max sleep times. > > - Ted > I think that it could be useful to have a tunable knob for the max sleep time. It would certainly be capped at a much lower level for a RAM disk than it would be for a very large RAID volume... Ric