对比Linux IO调度算法对顺序写的影响

操作系统信息

[root@test1 ~]# more /etc/redhat-release

Red Hat Enterprise Linux Server release 6.4 (Santiago)

[root@test1 ~]#

[root@test1 ~]# uname -a

Linux test1 2.6.32-358.el6.x86_64 #1 SMP Tue Jan 29 11:47:41 EST 2013 x86_64 x86_64 x86_64 GNU/Linux

[root@test1 ~]#

[root@test1 ~]# dme

dmesg         dmeventd      dmevent_tool

[root@test1 ~]# dmesg |grep scheduler

io scheduler noop registered

io scheduler anticipatory registered

io scheduler deadline registered

io scheduler cfq registered (default)

[root@test1 ~]# 

硬盘型号

[root@test1 ext4]# smartctl -a /dev/sdb

smartctl 5.43 2012-06-30 r3573 [x86_64-linux-2.6.32-358.el6.x86_64] (local build)

Copyright (C) 2002-12 by Bruce Allen,

=== START OF INFORMATION SECTION ===

Model Family:     Western Digital Caviar Blue Serial ATA

Device Model:     WDC WD3200AAKS-75L9A0

Serial Number:    WD-WMAV2F258203

LU WWN Device Id: 5 0014ee 0018fdbb9

Firmware Version: 02.03E02

User Capacity:    320,072,933,376 bytes [320 GB]

Sector Size:      512 bytes logical/physical

Device is:        In smartctl database [for details use: -P show]

ATA Version is:   8

ATA Standard is:  Exact ATA specification draft version not indicated

Local Time is:    Wed May  7 12:13:33 2014 CST

SMART support is: Available - device has SMART capability.

SMART support is: Enabled

文件系统 EXT4

测试脚本

time sysbench --num-threads=16 --file-block-size=8192 --test=fileio --file-total-size=5G --file-test-mode=seqwr prepare

time sysbench --num-threads=16 --file-block-size=8192 --test=fileio --file-total-size=5G --file-test-mode=seqwr run

time sysbench --num-threads=16 --file-block-size=8192 --test=fileio --file-total-size=5G --file-test-mode=seqwr cleanup

time sysbench --num-threads=16 --file-block-size=8192 --test=fileio --file-total-size=5G --file-test-mode=seqrewr prepare

time sysbench --num-threads=16 --file-block-size=8192 --test=fileio --file-total-size=5G --file-test-mode=seqrewr run

time sysbench --num-threads=16 --file-block-size=8192 --test=fileio --file-total-size=5G --file-test-mode=seqrewr cleanup

准备工作

[root@test1 ~]# umount /ext4/

[root@test1 ~]# mkfs.ext4 /dev/sdb

sdb   sdb1  sdb2  sdb3  sdb4  sdb5  sdb6  sdb7

[root@test1 ~]# mkfs.ext4 /dev/sdb2

mke2fs 1.41.12 (17-May-2010)

Filesystem label=

OS type: Linux

Block size=4096 (log=2)

Fragment size=4096 (log=2)

Stride=0 blocks, Stripe width=0 blocks

3278576 inodes, 13109040 blocks

655452 blocks (5.00%) reserved for the super user

First data block=0

Maximum filesystem blocks=4294967296

401 block groups

32768 blocks per group, 32768 fragments per group

8176 inodes per group

Superblock backups stored on blocks:

        32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,

        4096000, 7962624, 11239424

Writing inode tables: done

Creating journal (32768 blocks): done

Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 20 mounts or

180 days, whichever comes first.  Use tune2fs -c or -i to override.

[root@test1 ~]# tune2fs -l /dev/sdb2

tune2fs 1.41.12 (17-May-2010)

Filesystem volume name:  

Last mounted on:          

Filesystem UUID:          2b14ee13-d2bf-4928-8bc5-49c004a79f4b

Filesystem magic number:  0xEF53

Filesystem revision #:    1 (dynamic)

Filesystem features:      has_journal ext_attr resize_inode dir_index filetype extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize

Filesystem flags:         signed_directory_hash

Default mount options:    (none)

Filesystem state:         clean

Errors behavior:          Continue

Filesystem OS type:       Linux

Inode count:              3278576

Block count:              13109040

Reserved block count:     655452

Free blocks:              12857228

Free inodes:              3278565

First block:              0

Block size:               4096

Fragment size:            4096

Reserved GDT blocks:      1020

Blocks per group:         32768

Fragments per group:      32768

Inodes per group:         8176

Inode blocks per group:   511

Flex block group size:    16

Filesystem created:       Wed May  7 11:12:08 2014

Last mount time:          n/a

Last write time:          Wed May  7 11:12:20 2014

Mount count:              0

Maximum mount count:      20

Last checked:             Wed May  7 11:12:08 2014

Check interval:           15552000 (6 months)

Next check after:         Mon Nov  3 11:12:08 2014

Lifetime writes:          933 MB

Reserved blocks uid:      0 (user root)

Reserved blocks gid:      0 (group root)

First inode:              11

Inode size:               256

Required extra isize:     28

Desired extra isize:      28

Journal inode:            8

Default directory hash:   half_md4

Directory Hash Seed:      0ff44d76-fc7d-4b92-924b-6258a13c161e

Journal backup:           inode blocks

[root@test1 ~]#

[root@test1 ~]# vi /etc/fstab

[root@test1 ~]#

[root@test1 ~]# mount /ext4

[root@test1 ~]# more /proc/mounts |grep sdb2

/dev/sdb2 /ext4 ext4 rw,relatime,barrier=1,data=ordered 0 0

[root@test1 ~]#

看看当前的I/O 调度策略

[root@test1 ~]# more /sys/block/sdb/queue/scheduler

noop anticipatory deadline [cfq]

[root@test1 ~]#

测试顺序

cfq noop anticipatory deadline 

开始测试

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler

noop anticipatory deadline [cfq]

[root@test1 ext4]# echo cfq  > /sys/block/sdb/queue/scheduler

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler

noop anticipatory deadline [cfq]

[root@test1 ext4]#

执行测试脚本......

[root@test1 ext4]#

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler                          

noop anticipatory deadline [cfq]

[root@test1 ext4]# echo noop  > /sys/block/sdb/queue/scheduler                  

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler

[noop] anticipatory deadline cfq

[root@test1 ext4]#

执行测试脚本......

[root@test1 ext4]# echo anticipatory  > /sys/block/sdb/queue/scheduler

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler

noop [anticipatory] deadline cfq

[root@test1 ext4]#

执行测试脚本......

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler                          noop [anticipatory] deadline cfq

[root@test1 ext4]# echo deadline  > /sys/block/sdb/queue/scheduler

[root@test1 ext4]# more /sys/block/sdb/queue/scheduler

noop anticipatory [deadline] cfq

[root@test1 ext4]#

执行测试脚本......

结果对比

IO Scheduler	seqrw		seqrerw
	transferred /sec (Mb/sec)	Requests/sec	transferred /sec (Mb/sec)	Requests/sec
cfq	90.146	11538.69	77.983	9981.79
noop	91.315	11688.32	80.119	10255.28
anticipatory	90.484	11581.97	80.274	10275.05
deadline	93.242	11934.95	83.188	10648.09

IO Scheduler	seqrw	seqrerw
	transferred /sec (Mb/sec)	transferred /sec (Mb/sec)
cfq	90.146	77.983
noop	91.315	80.119
anticipatory	90.484	80.274
deadline	93.242	83.188

 

IO Scheduler	seqrw	seqrerw
	Requests/sec	Requests/sec
cfq	11538.69	9981.79
noop	11688.32	10255.28
anticipatory	11581.97	10275.05
deadline	11934.95	10648.09

 

11934.95-11538.69=396.26/11538.69=0.03434185

#end

---

附件列表