From mboxrd@z Thu Jan 1 00:00:00 1970 From: Loic Dachary Subject: Re: What crush ruleset for a given SHEC configuration ? Date: Tue, 19 May 2015 12:47:00 +0200 Message-ID: <555B14A4.5000008@dachary.org> References: <555ADA6F.7000702@dachary.org> <870DE8DBB716524BAE51B2D499EC81E40AAF9098@g01jpexmbyt24> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="257Bn8i7qMslu9wPvP6GDD2Uo2wCT5r9C" Return-path: Received: from mail2.dachary.org ([91.121.57.175]:37723 "EHLO smtp.dmail.dachary.org" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1753405AbbESKrE (ORCPT ); Tue, 19 May 2015 06:47:04 -0400 In-Reply-To: <870DE8DBB716524BAE51B2D499EC81E40AAF9098@g01jpexmbyt24> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: "Miyamae, Takeshi" Cc: Ceph Development This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --257Bn8i7qMslu9wPvP6GDD2Uo2wCT5r9C Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable On 19/05/2015 11:11, Miyamae, Takeshi wrote: > Hi Loic, >=20 >> to recover D6 which is in rack 2 it may be necessary to use P2 from ra= ck 1 >=20 > Because D6 can be recovered from P3 as well, P2 is not necessarily used= to recover D6. > However, when D6 is recovered from P3, D5 which is in rack 1 must be re= ad. > Therefore, I believe Ceph-LRC is more efficient when network resources = is poor such as > in case of Geo-Replication. SHEC should be used in a single data center= that has enough > network resources. Thanks for the detailed explanation, it is clear now :-) Cheers >=20 > Best regards, > Takeshi Miyamae >=20 > -----Original Message----- > From: Loic Dachary [mailto:loic@dachary.org]=20 > Sent: Tuesday, May 19, 2015 3:39 PM > To: Miyamae, Takeshi/=E5=AE=AE=E5=89=8D =E5=89=9B > Cc: Ceph Development > Subject: What crush ruleset for a given SHEC configuration ? >=20 > Hi Takeshi, >=20 > In the context of http://ceph.com/docs/master/rados/operations/erasure-= code-shec/ it would be useful to have a more detailed explanation of why = SHEC is more efficient during recovery (in the introduction).=20 >=20 > Am I correct to assume that SHEC does not provide a way to control the = locality of the chunks ? For instance in the following scenario: >=20 > rack 1 has 10 OSDs > rack 2 has 10 OSDs >=20 > a crush ruleset is made to provide 15 OSDs with 7 in the first rack, 8 = in the last rack: the first 7 are in rack 1, the last 8 in rack 2. When S= HEC is used with such a crush ruleset, it cannot guarantee that the loss = of one chunk in rack 2 can always be recovered with chunks from rack 2. W= hen reading at figure 3 of >=20 > https://wiki.ceph.com/Planning/Blueprints/Hammer/Shingled_Erasure_Code_= %28SHEC%29 >=20 > with D1 to D5, P1 and P2 in rack 1 and D6 to D10, P3, P4, P5 in rack 2,= my understanding is that to recover D6 which is in rack 2 it may be nece= ssary to use P2 from rack 1. And to recover D5 which is in rack 1 it may = be necessary to use P3 from rack 2. >=20 > Maybe I'm missing something ? Thanks in advance for your explanations := -) >=20 > Cheers >=20 --=20 Lo=C3=AFc Dachary, Artisan Logiciel Libre --257Bn8i7qMslu9wPvP6GDD2Uo2wCT5r9C 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) iEYEARECAAYFAlVbFKQACgkQ8dLMyEl6F20hbACgpBti3cuJ8W/FETyWZYrMcwKl C+sAn0+x/VUa7U35I2CQS4sz7U6NqIQa =hXyE -----END PGP SIGNATURE----- --257Bn8i7qMslu9wPvP6GDD2Uo2wCT5r9C--