From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dan Mick Subject: Re: rbd STDIN import does not work / wip-rbd-export-stdout Date: Sat, 01 Dec 2012 16:28:14 -0800 Message-ID: <50BAA09E.9070506@inktank.com> References: <50B36C9F.80107@profihost.ag> <50B36ED1.1020108@filoo.de> <50B37027.5000701@profihost.ag> <2EC6C33E-4909-4AE5-B8C0-5D7A5856EA57@profihost.ag> <50B3CA6F.6000000@inktank.com> <50B484FF.9070500@profihost.ag> <50BA6746.4010609@profihost.ag> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mail-da0-f46.google.com ([209.85.210.46]:37203 "EHLO mail-da0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753532Ab2LBA2b (ORCPT ); Sat, 1 Dec 2012 19:28:31 -0500 Received: by mail-da0-f46.google.com with SMTP id p5so772804dak.19 for ; Sat, 01 Dec 2012 16:28:31 -0800 (PST) In-Reply-To: <50BA6746.4010609@profihost.ag> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Stefan Priebe Cc: "ceph-devel@vger.kernel.org" On 12/1/2012 12:23 PM, Stefan Priebe wrote: > Hi Dan, > > Am 27.11.2012 10:16, schrieb Stefan Priebe - Profihost AG: >> Am 26.11.2012 21:00, schrieb Dan Mick: >>> It writes zeros; there's no way for it to know how many zeros are >>> coming. It could make a half-hearted attempt depending on its buffer >>> size and the amount of data the source is willing to buffer. > > Yes but why not skip at leas if the WHOLE BUFFER is zero and then use > the buffer size = stripe size? > > Right now a 40GB image which has just 3GB used rest is zeros due to > using fstrim before doing the export takes a long time to import. We > could speed this up a lot if we would skip parts with only zeros. Great minds! I'm actually working on that now; as you say it's not hard. > Greets, > Stefab