From mboxrd@z Thu Jan 1 00:00:00 1970 From: NeilBrown Subject: Re: [RFC] Process requests instead of bios to use a scheduler Date: Mon, 2 Jun 2014 20:20:50 +1000 Message-ID: <20140602202050.14903534@notabene.brown> References: <5385DECE.5060507@profitbricks.com> <20140602093258.22aa2c05@notabene.brown> <538C4938.6010704@profitbricks.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=PGP-SHA1; boundary="Sig_/dyqGMXdqm23gRjlrxpME99N"; protocol="application/pgp-signature" Return-path: In-Reply-To: <538C4938.6010704@profitbricks.com> Sender: linux-raid-owner@vger.kernel.org To: Sebastian Parschauer Cc: Linux RAID , Florian-Ewald =?UTF-8?B?TcO8?= =?UTF-8?B?bGxlcg==?= List-Id: linux-raid.ids --Sig_/dyqGMXdqm23gRjlrxpME99N Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable On Mon, 02 Jun 2014 11:51:52 +0200 Sebastian Parschauer wrote: > > Having a scheduler for RAID0 doesn't make any sense to me. > > RAID0 simply passes each request down to the appropriate underlying dev= ice. > > That device then does its own scheduling. > >=20 > > Adding a scheduler may well make sense for RAID1 (the current "schedule= r" > > only does some read balancing and is rather simplistic) and for RAID4/5= /6/10. > >=20 > > But not for RAID0 .... was that a typo? >=20 > Nope, we have our RAID-1+0. So it is more or less a RAID-10 and putting > the scheduler to this RAID-0 layer makes sense for us. I still cannot imagine how this would work. RAID-0 has no decisions to mak= e, so no where for a scheduler to fit. Just to clarify: is this md/raid0 over md/raid1 or md/raid0 over hardware/raid1? > > Could you do a graph? I like graphs :-) > > I can certainly seem something has changed here... >=20 > Sure, please find the graphs attached. I've converted it into percentage > so that number of bios can be compared to number of requests. Thanks. > >=20 > > Show me the code and I might be able to provide a more detailed opinion. >=20 > I would say let the user decide whether an MD device should be equipped > with a scheduler or not. We can port our code to latest kernel + latest > mdadm and send you a patch set for testing. Just give me some time to do = it. In the first instance, I just want to get a concrete idea of what you have done because what you have said doesn't make sense to me. I'm happy to look at code against a not-quite-current kernel to get that idea. But I'm also happy for it to be against the latest, whatever suits you. NeilBrown --Sig_/dyqGMXdqm23gRjlrxpME99N Content-Type: application/pgp-signature; name=signature.asc Content-Disposition: attachment; filename=signature.asc -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQIVAwUBU4xQAjnsnt1WYoG5AQKJZhAAncoIsuw6ErcQYUqWPXiVaGdkpdPZVJv3 6fUYQ30u1Cu84eBXQWOjtdPntLS0BiW8DumNuUNT4KWyoddJ/VfgYZOveacaCvZg nBNQkgWhEi4NKhXmxnJ/oa5uK8XkMWHTRU6AcqbsREVQKYocgf3gk+qW8uEGbI4w MfDDPwg3GkFszU1wQhuEM8nH49EuNxTTVL2CJYadexuOGFAIQHH3JZBlWq/uXfzs UrTiT0fdvesjaEaT2qoHm44AiHvy9108GG8gqANWfbL6+nzZ58LFg5TmR+xL4EEh sqm9ub8WzSjxkdbMO268QnQTqyMhspoj/UuexWt1s9JlBisxXCY5gdVmO0IvaDQX dZe4xja0a3Umh2lpn+XO0ObfH5vz55C7hwnxtVBka7Vop09f0xXKKRJ3QpPoRgGH Bu4yNKxS3Y7APsDFbNmNw0t3DaDfD7EVZ/yagGke5nkze2wXxoMqhzSjvSo1y01r Kj5F/Kjqn0WXE6Kpj0StciXAzwkBewiD5dpmLgLIQIvEat9KF56pBcByhf2kDgF8 Vr1Wvmo+ncUM6q5fs5t4XtDXnoVdx4ZjuQs8FgCs79i96OrCc8U6Ebi56Tqm+/Yd C7OElrE8k3ABgEXG/vclVVS9y9TXvAQEqEKXb/IfEWhyU3dnlxiV7b/mgqkE2HKK yuGBzy3WQXI= =GSOm -----END PGP SIGNATURE----- --Sig_/dyqGMXdqm23gRjlrxpME99N--