Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752083Ab0LMOaO (ORCPT ); Mon, 13 Dec 2010 09:30:14 -0500 Received: from mx1.redhat.com ([209.132.183.28]:19715 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751242Ab0LMOaM (ORCPT ); Mon, 13 Dec 2010 09:30:12 -0500 Date: Mon, 13 Dec 2010 09:29:57 -0500 From: Vivek Goyal To: Gui Jianfeng Cc: Jens Axboe , Corrado Zoccolo , Chad Talbott , Nauman Rafique , Divyesh Shah , linux kernel mailing list Subject: Re: [PATCH 0/8 v2] Introduce CFQ group hierarchical scheduling and "use_hierarchy" interface Message-ID: <20101213142957.GA20454@redhat.com> References: <4CDF7BC5.9080803@cn.fujitsu.com> <4CDF9CC6.2040106@cn.fujitsu.com> <20101115165319.GI30792@redhat.com> <4CE2718C.6010406@kernel.dk> <4D057A6A.8060000@cn.fujitsu.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4D057A6A.8060000@cn.fujitsu.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 20893 Lines: 456 On Mon, Dec 13, 2010 at 09:44:10AM +0800, Gui Jianfeng wrote: > Hi > > Previously, I posted a patchset to add support of CFQ group hierarchical scheduling > in the way that it puts all CFQ queues in a hidden group and schedules with other > CFQ group under their parent. The patchset is available here, > http://lkml.org/lkml/2010/8/30/30 > > Vivek think this approach isn't so instinct that we should treat CFQ queues > and groups at the same level. Here is the new approach for hierarchical > scheduling based on Vivek's suggestion. The most big change of CFQ is that > it gets rid of cfq_slice_offset logic, and makes use of vdisktime for CFQ > queue scheduling just like CFQ group does. But I still give cfqq some jump > in vdisktime based on ioprio, thanks for Vivek to point out this. Now CFQ > queue and CFQ group uses the same scheduling algorithm. Hi Gui, Thanks for the patches. Few thoughts. - I think we can implement vdisktime jump logic for both cfq queue and cfq groups. So any entity (queue/group) which is being backlogged fresh will get the vdisktime jump but anything which has been using its slice will get queued at the end of tree. - Have you done testing in true hierarchical mode. In the sense that create atleast two level of hierarchy and see if bandwidth division is happening properly. Something like as follows. root / \ test1 test2 / \ / \ G1 G2 G3 G4 - On what kind of storage you have been doing your testing? I have noticed that IO controllers works well only with idling on and with idling on performance is bad on high end storage. The simple reason being that an storage array can support multiple IOs at the same time and if we are idling on queue or group in an attempt to provide fairness it hurts. It hurts especially more if we are doing random IO (I am assuming this is more typical of workloads). So we need to come up with a proper logic so that we can provide some kind of fairness even with idle disabled. I think that's where this vdisktime jump logic comes into picture and is important to get it right. So can you also do some testing with idle disabled (both queue and group) and see if the vdisktime logic is helping with providing some kind of service differentation. I think results will vary based on what is the storage and what queue depth are you driving. You can even try to do this testing on an SSD. Thanks Vivek > > "use_hierarchy" interface is now added to switch between hierarchical mode > and flat mode. For this time being, this interface only appears in Root Cgroup. > > V1 -> V2 Changes: > - Raname "struct io_sched_entity" to "struct cfq_entity" and don't differentiate > queue_entity and group_entity, just use cfqe instead. > - Give newly added cfqq a small vdisktime jump accord to its ioprio. > - Make flat mode as default CFQ group scheduling mode. > - Introduce "use_hierarchy" interface. > - Update blkio cgroup documents > > [PATCH 1/8 v2] cfq-iosched: Introduce cfq_entity for CFQ queue > [PATCH 2/8 v2] cfq-iosched: Introduce cfq_entity for CFQ group > [PATCH 3/8 v2] cfq-iosched: Introduce vdisktime and io weight for CFQ queue > [PATCH 4/8 v2] cfq-iosched: Extract some common code of service tree handling for CFQ queue and CFQ group > [PATCH 5/8 v2] cfq-iosched: Introduce hierarchical scheduling with CFQ queue and group at the same level > [PATCH 6/8] blkio-cgroup: "use_hierarchy" interface without any functionality > [PATCH 7/8] cfq-iosched: Add flat mode and switch between two modes by "use_hierarchy" > [PATCH 8/8] blkio-cgroup: Document for blkio.use_hierarchy. > > Benchmarks: > I made use of Vivek's iostest to perform some benchmarks on my box. I tested different workloads, and > didn't see any performance drop comparing to vanilla kernel. The attached files are some performance > numbers on vanilla Kernel, patched kernel with flat mode and patched kernel with hierarchical mode. > > > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=brr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[brr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > brr 1 1 294 692 1101 1526 3613 > brr 1 2 176 420 755 1281 2632 > brr 1 4 160 323 583 1253 2319 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=brr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[brr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > brr 1 1 380 738 1092 1439 3649 > brr 1 2 171 413 733 1242 2559 > brr 1 4 188 350 665 1193 2396 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=bsr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[bsr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > bsr 1 1 6856 11480 17644 22647 58627 > bsr 1 2 2592 5409 8464 13300 29765 > bsr 1 4 2502 4635 7640 12909 27686 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=bsr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[bsr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > bsr 1 1 6913 11643 17843 22909 59308 > bsr 1 2 6682 11234 15527 19410 52853 > bsr 1 4 5209 10882 15002 18167 49260 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drr 1 1 298 701 1117 1538 3654 > drr 1 2 190 372 731 1244 2537 > drr 1 4 147 322 563 1143 2175 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drr 1 1 370 713 1050 1416 3549 > drr 1 2 192 434 755 1265 2646 > drr 1 4 157 333 677 1159 2326 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drw iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drw] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drw 1 1 595 1272 2007 2737 6611 > drw 1 2 269 690 1407 1953 4319 > drw 1 4 145 396 978 1752 3271 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drw iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drw] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drw 1 1 604 1310 1827 2778 6519 > drw 1 2 287 723 1368 1887 4265 > drw 1 4 170 407 979 1575 3131 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=brr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[brr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > brr 1 1 287 690 1096 1506 3579 > brr 1 2 189 404 800 1283 2676 > brr 1 4 141 317 557 1106 2121 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=brr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[brr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > brr 1 1 386 715 1071 1437 3609 > brr 1 2 187 401 717 1258 2563 > brr 1 4 296 767 1553 32 2648 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=bsr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[bsr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > bsr 1 1 6971 11459 17789 23158 59377 > bsr 1 2 2592 5536 8679 13389 30196 > bsr 1 4 2194 4635 7820 13984 28633 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=bsr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[bsr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > bsr 1 1 6851 11588 17459 22297 58195 > bsr 1 2 6814 11534 16141 20426 54915 > bsr 1 4 5118 10741 13994 17661 47514 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drr 1 1 297 689 1097 1522 3605 > drr 1 2 175 426 757 1277 2635 > drr 1 4 150 330 604 1100 2184 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drr 1 1 379 735 1077 1436 3627 > drr 1 2 190 404 760 1247 2601 > drr 1 4 155 333 692 1044 2224 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drw iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drw] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drw 1 1 566 1293 2001 2686 6546 > drw 1 2 225 662 1233 1902 4022 > drw 1 4 147 379 922 1761 3209 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drw iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drw] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drw 1 1 579 1226 2020 2823 6648 > drw 1 2 276 689 1288 2068 4321 > drw 1 4 183 399 798 2113 3493 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=brr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[brr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > brr 1 1 289 684 1098 1508 3579 > brr 1 2 178 421 765 1228 2592 > brr 1 4 144 301 585 1094 2124 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=brr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[brr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > brr 1 1 375 734 1081 1434 3624 > brr 1 2 172 397 700 1201 2470 > brr 1 4 154 316 573 1087 2130 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=bsr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[bsr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > bsr 1 1 6818 11510 17820 23239 59387 > bsr 1 2 2643 5502 8728 13329 30202 > bsr 1 4 2166 4785 7344 12954 27249 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=bsr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[bsr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > bsr 1 1 6979 11629 17782 23064 59454 > bsr 1 2 6803 11274 15865 20024 53966 > bsr 1 4 5292 10674 14504 17674 48144 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drr 1 1 298 694 1116 1540 3648 > drr 1 2 183 405 721 1197 2506 > drr 1 4 151 296 553 1119 2119 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drr iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drr] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drr 1 1 367 724 1078 1433 3602 > drr 1 2 184 418 744 1245 2591 > drr 1 4 157 295 562 1122 2136 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drw iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=0 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drw] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drw 1 1 582 1271 1948 2754 6555 > drw 1 2 277 700 1294 1970 4241 > drw 1 4 172 345 928 1585 3030 > > > Host=localhost.localdomain Kernel=2.6.37-rc2-Block-+ > GROUPMODE=1 NRGRP=4 > DIR=/mnt/iostestmnt/fio DEV=/dev/sdb2 > Workload=drw iosched=cfq Filesz=512M bs=32k > group_isolation=1 slice_idle=8 group_idle=8 quantum=8 > ========================================================================= > AVERAGE[drw] [bw in KB/s] > ------- > job Set NR cgrp1 cgrp2 cgrp3 cgrp4 total > --- --- -- ----------------------------------- > drw 1 1 586 1296 1888 2739 6509 > drw 1 2 294 749 1360 1931 4334 > drw 1 4 156 337 814 1806 3113 > > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/