From mboxrd@z Thu Jan 1 00:00:00 1970 From: Loic Dachary Subject: Crush ruleset for locally repairable codes (LRC) Date: Sun, 08 Jun 2014 16:30:17 +0200 Message-ID: <53947379.1020407@dachary.org> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="1uLSm4lA7J7hdmdCA7DouJBn9pHvHsoN6" Return-path: Received: from smtp.dmail.dachary.org ([91.121.254.229]:37659 "EHLO smtp.dmail.dachary.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753301AbaFHOaU (ORCPT ); Sun, 8 Jun 2014 10:30:20 -0400 Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Andreas-Joachim Peters Cc: Ceph Development This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --1uLSm4lA7J7hdmdCA7DouJBn9pHvHsoN6 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Hi Andreas, The draft ruleset creation matching for LRC is documented at https://github.com/dachary/ceph/commit/2ae645438243232392c11fd80fd960f622= 1fd2a0#diff-5518964bc98a094a784ce2d17a5b0cc1R147 tested at https://github.com/dachary/ceph/commit/2ae645438243232392c11fd80fd960f622= 1fd2a0#diff-300a5998686e416e9255777f82fe8c57R90 and implemented at https://github.com/dachary/ceph/commit/2ae645438243232392c11fd80fd960f622= 1fd2a0#diff-e8889adeda9017a36ae4f5000b9c2534R42 There are two points for which your advice would be most appreciated: * The default ruleset is the same as for jerasure, i.e. host is the failu= re domain. The primary reason is that it is likely to work anywhere, henc= e it is a good default. Although it does not match any actual use case, i= t there is a side effect that can be observed externally (i.e. reduced in= ter-host bandwidth during recovery) and that matches the expected benefit= =2E It would be an actual benefit if machines were connected to a hub but= that is unlikely to be the case nowadays ;-) * Control of the crush ruleset steps is made explicit and disconnected fr= om the description of the layers. Trying to link the encoding steps to th= e crush description creates confusion because the logic is different. The= crush ruleset is designed first, thinking about failure domains. The era= sure code processing is designed next, using the N OSDs provided by crush= =2E Hopefully the examples and logic documented at https://github.com/dachary/ceph/commit/2ae645438243232392c11fd80fd960f622= 1fd2a0#diff-5518964bc98a094a784ce2d17a5b0cc1R20 make sense to the Ceph operator.=20 Cheers --=20 Lo=EFc Dachary, Artisan Logiciel Libre --1uLSm4lA7J7hdmdCA7DouJBn9pHvHsoN6 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlOUc3kACgkQ8dLMyEl6F22X9wCcD5DKyIPYBLOUTrEpjt5iYeW7 tjAAoKMDYnhpjpU/UxlaiV5mTz8Fzc+P =fV1T -----END PGP SIGNATURE----- --1uLSm4lA7J7hdmdCA7DouJBn9pHvHsoN6--