前言

sysbench是一个开源的、模块化的、跨平台的多线程性能测试工具,可以用来进行CPU、内存、磁盘I/O、线程、数据库的性能测试。sysbench是基于LuaJIT的可编写脚本的多线程基准测试工具。

sysbench附带以下捆绑的基准测试:

  • oltp_*.lua:类似OLTP的数据库基准测试的集合
  • fileio:文件系统级基准
  • cpu:简单的CPU基准测试
  • memory:内存访问基准
  • threads:基于线程的调度程序基准
  • mutex:POSIX互斥基准测试

特征

  • 提供有关速率和延迟的大量统计信息,包括延迟百分位数和直方图;
  • 即使有数千个并发线程,开销也很低。sysbench每秒能够生成和跟踪数亿个事件。
  • 通过在用户提供的Lua脚本中实现预定义的挂钩,可以轻松创建新的基准测试;
  • 可以也可以作为一个通用的Lua解释,只需更换#!/usr/bin/lua#!/usr/bin/sysbench在你的脚本。

安装

安装以Centos为例,其他版本或者mac请参考sysbench说明

1.下载安转

wget https://github.com/akopytov/sysbench/archive/1.0.zip -O "sysbench-1.0.zip"

unzip sysbench-1.0.zip

cd sysbench-1.0

//安装依赖
yum install -y automake libtool
 

2.编译安装

./autogen.sh
./configure
export LD_LIBRARY_PATH=/usr/include/mysql
make
make install

3.监测是否安装成功

[root@guanbin-k8s-master ~]# sysbench --version
sysbench 1.0.19  

测试

1. cpu测试

cpu测试主要是进行素数的加法运算,我们的例子中,指定了最大的质数发生器数量为 20000

[root@guanbin-k8s-master ~]# sysbench --test=cpu --cpu-max-prime=20000 run
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any options.
sysbench 1.0.19 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options:
Number of threads: 1
Initializing random number generator from current time Prime numbers limit: 20000 Initializing worker threads... Threads started! CPU speed:
events per second: 168.59 General statistics:
total time: 10.0027s
total number of events: 1687 Latency (ms):
min: 5.63
avg: 5.93
max: 66.09
95th percentile: 6.43
sum: 9999.10 Threads fairness:
events (avg/stddev): 1687.0000/0.00
execution time (avg/stddev): 9.9991/0.00

执行时间 大约为10秒(这个测试最好是两台机器进行对比测试,才能发现那个机器有问题)

2.对内存(memory)进行测试 
   测试是在内存中传输 10G 的数据量,每个 block 大小为 8K
[root@guanbin-k8s-master ~]# sysbench --test=memory --memory-block-size=8k --memory-total-size=10G run
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any options.
sysbench 1.0.19 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options:
Number of threads: 1
Initializing random number generator from current time Running memory speed test with the following options:
block size: 8KiB
total size: 10240MiB
operation: write
scope: global Initializing worker threads... Threads started! Total operations: 1310720 (724931.70 per second) 10240.00 MiB transferred (5663.53 MiB/sec) General statistics:
total time: 1.8045s
total number of events: 1310720 Latency (ms):
min: 0.00
avg: 0.00
max: 12.44
95th percentile: 0.00
sum: 1441.61 Threads fairness:
events (avg/stddev): 1310720.0000/0.00
execution time (avg/stddev): 1.4416/0.00

totaltime为1.8045s

3.对disk 的IO测试

新建一个临时目录,并确保磁盘剩余空间足够(--file-total-size不要超过剩余容量)

测试场景:16个线程,创建128个文件,总共占用磁盘空间10GB,测试磁盘读写效率

运行结束后要删除临时测试文件: rm -rf test_file.*

root@guanbin-k8s-master tmp]# sysbench --test=fileio --num-threads=16 --file-total-size=1G --file-test-mode=rndrw prepare
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any options.
WARNING: --num-threads is deprecated, use --threads instead
sysbench 1.0.19 (using bundled LuaJIT 2.1.0-beta2) 128 files, 8192Kb each, 1024Mb total
Creating files for the test...
Extra file open flags: (none)
Creating file test_file.0
Creating file test_file.1
Creating file test_file.2
Creating file test_file.3
Creating file test_file.4
Creating file test_file.5
Creating file test_file.6
Creating file test_file.7
Creating file test_file.8
Creating file test_file.9
Creating file test_file.10
Creating file test_file.11
Creating file test_file.12
Creating file test_file.13
Creating file test_file.14
Creating file test_file.15
Creating file test_file.16
Creating file test_file.17
Creating file test_file.18
Creating file test_file.19
Creating file test_file.20
Creating file test_file.21
Creating file test_file.22
Creating file test_file.23
Creating file test_file.24
Creating file test_file.25
Creating file test_file.26
Creating file test_file.27
Creating file test_file.28
Creating file test_file.29
Creating file test_file.30
Creating file test_file.31
Creating file test_file.32
Creating file test_file.33
Creating file test_file.34
Creating file test_file.35
Creating file test_file.36
Creating file test_file.37
Creating file test_file.38
Creating file test_file.39
Creating file test_file.40
Creating file test_file.41
Creating file test_file.42
Creating file test_file.43
Creating file test_file.44
Creating file test_file.45
Creating file test_file.46
Creating file test_file.47
Creating file test_file.48
Creating file test_file.49
Creating file test_file.50
Creating file test_file.51
Creating file test_file.52
Creating file test_file.53
Creating file test_file.54
Creating file test_file.55
Creating file test_file.56
Creating file test_file.57
Creating file test_file.58
Creating file test_file.59
Creating file test_file.60
Creating file test_file.61
Creating file test_file.62
Creating file test_file.63
Creating file test_file.64
Creating file test_file.65
Creating file test_file.66
Creating file test_file.67
Creating file test_file.68
Creating file test_file.69
Creating file test_file.70
Creating file test_file.71
Creating file test_file.72
Creating file test_file.73
Creating file test_file.74
Creating file test_file.75
Creating file test_file.76
Creating file test_file.77
Creating file test_file.78
Creating file test_file.79
Creating file test_file.80
Creating file test_file.81
Creating file test_file.82
Creating file test_file.83
Creating file test_file.84
Creating file test_file.85
Creating file test_file.86
Creating file test_file.87
Creating file test_file.88
Creating file test_file.89
Creating file test_file.90
Creating file test_file.91
Creating file test_file.92
Creating file test_file.93
Creating file test_file.94
Creating file test_file.95
Creating file test_file.96
Creating file test_file.97
Creating file test_file.98
Creating file test_file.99
Creating file test_file.100
Creating file test_file.101
Creating file test_file.102
Creating file test_file.103
Creating file test_file.104
Creating file test_file.105
Creating file test_file.106
Creating file test_file.107
Creating file test_file.108
Creating file test_file.109
Creating file test_file.110
Creating file test_file.111
Creating file test_file.112
Creating file test_file.113
Creating file test_file.114
Creating file test_file.115
Creating file test_file.116
Creating file test_file.117
Creating file test_file.118
Creating file test_file.119
Creating file test_file.120
Creating file test_file.121
Creating file test_file.122
Creating file test_file.123
Creating file test_file.124
Creating file test_file.125
Creating file test_file.126
Creating file test_file.127
1073741824 bytes written in 110.77 seconds (9.24 MiB/sec).
[root@guanbin-k8s-master tmp]# sysbench --test=fileio --num-threads=16 --file-total-size=1G --file-test-mode=rndrw run
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any options.
WARNING: --num-threads is deprecated, use --threads instead
sysbench 1.0.19 (using bundled LuaJIT 2.1.0-beta2) Running the test with following options:
Number of threads: 16
Initializing random number generator from current time Extra file open flags: (none)
128 files, 8MiB each
1GiB total file size
Block size 16KiB
Number of IO requests: 0
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 100 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random r/w test
Initializing worker threads... Threads started! File operations:
reads/s: 157.44
writes/s: 104.96
fsyncs/s: 480.79 Throughput:
read, MiB/s: 2.46
written, MiB/s: 1.64 General statistics:
total time: 13.3355s
total number of events: 7865 Latency (ms):
min: 0.00
avg: 20.61
max: 1115.85
95th percentile: 142.39
sum: 162071.40 Threads fairness:
events (avg/stddev): 491.5625/141.20
execution time (avg/stddev): 10.1295/0.03  
吞吐量:读2.46MiB/s,写1.64MiB/s; 95%延迟 142.39ms
 
4.测试mysql性能
       测试时使用的脚本为lua脚本,可以使用sysbench自带脚本,也可以自己开发。对于大多数应用,使用sysbench自带的脚本就足够了。不同版本的sysbench中,lua脚本的位置可能不同,可以自己在sysbench路径下使用find命令搜索oltp.lua。
     P.S.:大多数数据服务都是oltp类型的,如果你不了解什么是oltp,那么大概率你的数据服务就是oltp类型的。MySQL OLTP(On-Line Transaction Processing联机事务处理过程);
  • (1)尽量不要在MySQL服务器运行的机器上进行测试,一方面可能无法体现网络(哪怕是局域网)的影响,另一方面,sysbench的运行(尤其是设置的并发数较高时)会影响MySQL服务器的表现。
  • (2)可以逐步增加客户端的并发连接数(--thread参数),观察在连接数不同情况下,MySQL服务器的表现;如分别设置为10,20,50,100等。
  • (3)一般执行模式选择complex即可,如果需要特别测试服务器只读性能,或不使用事务时的性能,可以选择simple模式或nontrx模式。
  • (4)如果连续进行多次测试,注意确保之前测试的数据已经被清理干净。

下面是sysbench使用的一个例子:

 注意:先准备数据库

   mysql> create database sbtest;

其中,执行模式为complex,使用了10个表,每个表有10万条数据,客户端的并发线程数为10,执行时间为120秒,每10秒生成一次报告。

sysbench ./tests/include/oltp_legacy/oltp.lua --mysql-host={yourhostip} --mysql-port=3306 --mysql-user={username} --mysql-password={password} --oltp-test-mode=complex --mysql-table-engine=innodb --oltp-tables-count=10 --oltp-table-size=100000 --threads=10 --time=120 --report-interval=10 prepare

其中的hostip、username、password请自行切换为自己真实的环境数据

查看执行结果

sysbench ./tests/include/oltp_legacy/oltp.lua --mysql-host={yourhostip} --mysql-port=3306 --mysql-user={username} --mysql-password={password} --oltp-test-mode=complex --oltp-tables-count=10 --oltp-table-size=100000 --threads=10 --time=120 --report-interval=10 run
Running the test with following options:
Number of threads: 10
Report intermediate results every 10 second(s)
Initializing random number generator from current time Initializing worker threads... Threads started! [ 10s ] thds: 10 tps: 17.19 qps: 362.78 (r/w/o: 254.64/72.76/35.38) lat (ms,95%): 1589.90 err/s: 0.00 reconn/s: 0.00
[ 20s ] thds: 10 tps: 33.50 qps: 670.08 (r/w/o: 469.06/134.02/67.01) lat (ms,95%): 530.08 err/s: 0.00 reconn/s: 0.00
[ 30s ] thds: 10 tps: 39.30 qps: 786.00 (r/w/o: 550.20/157.20/78.60) lat (ms,95%): 458.96 err/s: 0.00 reconn/s: 0.00
[ 40s ] thds: 10 tps: 38.80 qps: 775.45 (r/w/o: 543.16/154.69/77.59) lat (ms,95%): 634.66 err/s: 0.00 reconn/s: 0.00
[ 50s ] thds: 10 tps: 30.10 qps: 602.54 (r/w/o: 421.43/120.91/60.20) lat (ms,95%): 590.56 err/s: 0.00 reconn/s: 0.00
[ 60s ] thds: 10 tps: 33.90 qps: 678.00 (r/w/o: 474.60/135.60/67.80) lat (ms,95%): 419.45 err/s: 0.00 reconn/s: 0.00
[ 70s ] thds: 10 tps: 29.50 qps: 589.98 (r/w/o: 412.99/118.00/59.00) lat (ms,95%): 787.74 err/s: 0.00 reconn/s: 0.00
[ 80s ] thds: 10 tps: 39.90 qps: 798.02 (r/w/o: 558.62/159.60/79.80) lat (ms,95%): 569.67 err/s: 0.00 reconn/s: 0.00
[ 90s ] thds: 10 tps: 38.10 qps: 761.99 (r/w/o: 533.40/152.40/76.20) lat (ms,95%): 484.44 err/s: 0.00 reconn/s: 0.00
[ 100s ] thds: 10 tps: 43.90 qps: 877.99 (r/w/o: 614.60/175.60/87.80) lat (ms,95%): 383.33 err/s: 0.00 reconn/s: 0.00
[ 110s ] thds: 10 tps: 37.80 qps: 756.00 (r/w/o: 529.20/151.20/75.60) lat (ms,95%): 530.08 err/s: 0.00 reconn/s: 0.00
[ 120s ] thds: 10 tps: 44.40 qps: 888.01 (r/w/o: 621.60/177.60/88.80) lat (ms,95%): 419.45 err/s: 0.00 reconn/s: 0.00
SQL statistics:
queries performed:
read: 59836
write: 17096
other: 8548
total: 85480
transactions: 4274 (35.59 per sec.)
queries: 85480 (711.77 per sec.)
ignored errors: 0 (0.00 per sec.)
reconnects: 0 (0.00 per sec.) General statistics:
total time: 120.0919s
total number of events: 4274 Latency (ms):
min: 50.07
avg: 280.96
max: 2486.17
95th percentile: 559.50
sum: 1200810.10 Threads fairness:
events (avg/stddev): 427.4000/4.20
execution time (avg/stddev): 120.0810/0.01

其中,对于我们比较重要的信息包括:

queries:查询总数及qps

transactions:事务总数及tps

Latency-95th percentile:前95%的请求的最大响应时间,本例中是344毫秒,这个延迟非常大,是因为我用的MySQL服务器性能很差;在正式环境中这个数值是绝对不能接受的。

最后清理数据:

sysbench ./tests/include/oltp_legacy/oltp.lua --mysql-host={yourhostip} --mysql-port=3306 --mysql-user={username} --mysql-password={password} --oltp-tables-count=10  cleanup

清理的日志输出

sysbench 1.0.19 (using bundled LuaJIT 2.1.0-beta2)

Dropping table 'sbtest1'...
Dropping table 'sbtest2'...
Dropping table 'sbtest3'...
Dropping table 'sbtest4'...
Dropping table 'sbtest5'...
Dropping table 'sbtest6'...
Dropping table 'sbtest7'...
Dropping table 'sbtest8'...
Dropping table 'sbtest9'...
Dropping table 'sbtest10'...

服务器性能测试利器之sysbench的更多相关文章

  1. 后台任务利器之Hangfire

    后台任务利器之Hangfire 一.简述 Hangfire作为一款高人气且容易上手的分布式后台执行服务,支持多种数据库.在.net core的环境中,由Core自带的DI管理着生命周期,免去了在NF4 ...

  2. .NET Core开源组件:后台任务利器之Hangfire 转载 https://www.cnblogs.com/chenug/p/6655636.html

    .NET Core开源组件:后台任务利器之Hangfire   一.简述 Hangfire作为一款高人气且容易上手的分布式后台执行服务,支持多种数据库.在.net core的环境中,由Core自带的D ...

  3. 破解浏览器同源政策利器之JSONP

    本文是在了解了浏览器的同源规则之后,学习了破解这个规则的一个简单有效的方法->JSONP.主要通过阮一峰老师的博客学习 浏览器的同源规则 有这样一个背景,如果你通过银行的网站进行的取钱的交易,而 ...

  4. Win下必备利器之Cmder

    诚言,对于开发码字者,Mac和Linux果断要比Windows更贴心;但只要折腾下,Windows下也是有不少利器的.之前就有在Windows下效率必备软件一文中对此做了下记载:其虽没oh-my-zs ...

  5. C#编程利器之二:结构与枚举(Structure and enumeration)【转】

    C#编程利器之二:结构与枚举(Structure and enumeration) 在上一篇文章中,介绍了类如何封装程序中的对象.而实际中,出了类可以封装对象外,结构和枚举也可以封装一些对象,本文将着 ...

  6. 大数据并行计算利器之MPI/OpenMP

    大数据集群计算利器之MPI/OpenMP ---以连通域标记算法并行化为例 1 背景 图像连通域标记算法是从一幅栅格图像(通常为二值图像)中,将互相邻接(4邻接或8邻接)的具有非背景值的像素集合提取出 ...

  7. 新一代服务器性能测试工具Gatling

    新一代服务器性能测试工具Gatlinghttp://automationqa.com/forum.php?mod=viewthread&tid=2898&fromuid=2

  8. Now直播应用的后台服务器性能测试实践

    版权声明:本文由Oliver原创文章,转载请注明出处: 文章原文链接:https://www.qcloud.com/community/article/208 来源:腾云阁 https://www.q ...

  9. Loadrunner脚本开发-基于HTTP协议的流媒体视频在线播放服务器性能测试

    脚本开发-基于HTTP协议的流媒体视频在线播放服务器性能测试 by:授客 QQ:1033553122   目的 实现基于http协议的流媒体在线视频播放,服务器性能测试脚本,模拟用户浏览器方式在线播放 ...

随机推荐

  1. Inverse/Implicit Function Theorem

    目录 4.1 The Inverse Function Theorem The Implicit Function Theorem 4.3 Curves and Surfaces 4.4 The Mo ...

  2. 什么是NaN?它的类型是什么?如何可靠的测试一个值是否等于NaN?

    NaN属性表示"不是数字"的值.这个特殊值是由于一个操作数是非数字的(例如"abc"/4)或者因为操作的结果是非数字而无法执行的. 虽然看起来很简单,但是NaN ...

  3. 编写Java程序,模拟网上商城购物,当用户选好物品提交订单时,每笔订单会自动生成一个唯一的订单编号。

    查看本章节 查看作业目录 需求说明: 模拟网上商城购物,当用户选好物品提交订单时,每笔订单会自动生成一个唯一的订单编号.而部分电子商务网站在数据高峰期时,一毫秒可能需要处理近千笔的订单 现在简单模拟 ...

  4. Word批量设置表格自动调整

    1.说明 通过使用Word的宏功能, 批量设置表格, 根据窗口自动调整表格, 使所有表格的宽度和窗口一样, 而不用一个一个手动调整表格. 宏是一个批量处理程序命令, 正确地运用它可以提高工作效率. 微 ...

  5. C#中的显式转换

    大多数编程语言都支持显示转换,也称为强制转换,它与隐式转换相呼应,比如,一般的,整型可以通过隐式转换成浮点型,而浮点型需要通过强制转换成整型: int i = 32; double d = i;//整 ...

  6. Ranger开源流水线docker化实践案例

    1.背景 开发部门决定在Apache Ranger开源社区贡献代码,目标是个人国内排名Top1,世界排名Top2,并且在已经成为Ranger项目的Committer情况下,争取成为Ranger项目的P ...

  7. js 表格操作 全选和反选效果 案例

    <!DOCTYPE html> <html> <head lang="en"> <meta charset="UTF-8&quo ...

  8. react中使用charles实现本地数据mock

    首先下载charles软件地址,更详细的使用方法都包含在操作文档里,包含汉化版补丁(下载后查看) 链接:https://pan.baidu.com/s/1Q5rMbcX0Wus7AwdGUWa-Wg ...

  9. 安装Cacti-plugin

    安装pluginunzip cacti-plugin-0.8.7e-PA-v2.6.zip -d cacti-plugin-archcp -R cacti-plugin-arch/* /data/ww ...

  10. Git 的基本命令的使用

    1.获得Git仓库(克隆一份代码到本地仓库) git clone url 2.更新本地的代码 git pull 3.查看本地修改的文件 git status 4.将本地的修改加到stage中 git ...