From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ishtar.tlinx.org ([173.164.175.65]:55848 "EHLO Ishtar.sc.tlinx.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753180AbcKDHHm (ORCPT ); Fri, 4 Nov 2016 03:07:42 -0400 Message-ID: <581C310E.8080801@tlinx.org> Date: Thu, 03 Nov 2016 23:56:14 -0700 From: "L.A. Walsh" MIME-Version: 1.0 Subject: Re: [rfe]: finobt option separable from crc option? (was [rfc] larger batches for crc32c) References: <20161028031747.68472ac7@roar.ozlabs.ibm.com> <20161027214244.GO14023@dastard> <20161028131234.24a5cb6f@roar.ozlabs.ibm.com> <20161028160218.1af40906@roar.ozlabs.ibm.com> <20161031030853.GK22126@dastard> <20161101143918.4f154154@roar.ozlabs.ibm.com> <20161101054725.GZ14023@dastard> <58194CF8.1000501@tlinx.org> <20161103082950.GJ9920@dastard> <581B601A.4060403@tlinx.org> <20161103230018.GC28177@dastard> In-Reply-To: <20161103230018.GC28177@dastard> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Dave Chinner Cc: linux-xfs@vger.kernel.org Dave Chinner wrote: > If I had a dollar for every time someone said "hardware raid > protects me" I'd have retired years ago. ---- It's not a sole line of protection, but it does a fair job of *detection*. > Media scrubbing does not protect against misdirected writes, > corruptions to/from the storage, memory errors, software bugs, bad > compilers (yes, we've already had XFS CRCs find a compiler bug), > etc. --- Crc's don't _protect_ against such things either -- they can allow detection, but for protection ... I make due with xfsdump. As for crc's -- I already had the feature prevent me from creating and using new partitions where it didn't like me setting the disk UUID when creating a new partition -- something that I heard would be fixed -- but something that prevented me from testing the new finobt feature I was interested in. I'm sure I'm not alone in wanting to test in my environment, but _both_ with and without the crc option. I seem to remember, recently, that it took other kernel developers to disable xfs panicing the entire system on problem detection, with the idea of only disabling what was broken. Along the same lines, if I am using crc's and badness is detected, I'd want to be able to keep the disk online, though in "read-only" mode to allow me to explore what was wrong and find the best solution. > You may not care about this, but plenty of other XFS users do. --- That's a crock -- I never said I didn't care about having it -- just that I wanted to be able to run finobt with crc's being disabled. I noted after my last response, that you said your crc testing only jumped around the tests to compare crc's -- which sounded like they were still being generated but not checked -- no wonder the no-crc option was the slowest of tested options. > I can't do this for you. I don't know what your workload is, nor > what hardware you use. *Give me numbers* that I can work with - > something we can measure and fix. You need to do the work to show > there's an issue - I can't do that for you, and no amount of > demanding that I do will change that. I don't want to bother others with my testing, I just want the platform to be open enough for me to quietly do my own testing and go forward from there. I may be ignorant, but I'm also picky about things I care about -- thus I prefer to do _some_ things myself, thank-you. I realize that closed-platforms where the undesirable is bundled with the desirable and where end-users have no choice is the wave of the future. I realize that in many cases, large corporations are pushing these changes. Microsoft's last "open-update" that allowed selecting what updates you wanted is history and now you are forced to take everything or nothing. Such offerings are not made for the benefit of the users -- it's not something they want. But it doesn't matter, it's where large companies are pushing things so consumers will be easier to control. I wouldn't be surprised if this feature bundling was being pushed by project sponsors and that developers had little choice and I may not either, as I don't have the smallest fraction of the time I would need to add in options or changes to the SW I use, just to hold the "status quo", let alone make improvements. All I can usually do is ask and later, "re-ask" -- since policies in effect at one point in time may not be present later. Cheers! -l