From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from 0122700014.0.fullrate.dk ([95.166.99.235]:34402 "EHLO kernel.dk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752394Ab0F3He1 (ORCPT ); Wed, 30 Jun 2010 03:34:27 -0400 Message-ID: <4C2AF382.7000708@fusionio.com> Date: Wed, 30 Jun 2010 09:34:26 +0200 From: Jens Axboe MIME-Version: 1.0 Subject: Re: Have we changed number of fields in fio --minimal output References: <20100629193247.GA3819@redhat.com> <4C2AF2EC.1090806@fusionio.com> In-Reply-To: <4C2AF2EC.1090806@fusionio.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: fio-owner@vger.kernel.org List-Id: fio@vger.kernel.org To: Vivek Goyal Cc: fio@vger.kernel.org On 2010-06-30 09:31, Jens Axboe wrote: > On 2010-06-29 21:32, Vivek Goyal wrote: >> Hi, >> >> I was running latest fio and noticed that number of fields in fio >> --minimal output have gone up from 69 to 77. A increase of 8 fields. Don't >> see any update in --minimal documentation. Is it regarding total latency >> thing? > > Woops yes, there's a total latency in there as well now. Should just be > 4 extra fields, though. It gets logged after completion latency, but > before bandwidth stats. I'll update the documentation. > > Should we perhaps put a versioning field in there? Now would seem to be > a good time, since the output has changed anyway. I'm open to > suggestions from you or other terse output users. How about redesigning it a bit to make it more bullet proof... We could prefix series of fields with the value they are logging. So for instance, the 4 completion latency fields would include a clat prefix first: clat[%lu;%lu;%f;%f],foo[%lu;%lu],etc Would that not be more resilient to future changes? New fields would not bother you, and reordering should also be fine. Any other ideas? -- Jens Axboe