Windows 7环境下网站性能测试小工具 Apache Bench 和 Webbench使用和下载

1简要说明

Apache Bench Apache的网站性能测试小程序,Windows平台下的程序名简称ab.exe,要想获得这个80k的可执行程序,用户需要下载整个Apache Httpd软件包!还需要将其安装到电脑上,确实比较麻烦。

Webbench也是一款网站性能测试小程序,可以获得网站的吞吐率、传送速度等基本性能指标。该程序是在Linux下编程的,在Linux下编译安装使用都很方便,但在Windows 下的程序却没有。

为了方便起见,我将ab.exe 小程序单独拷贝出来,并且下载了webbench的源码,使用CodeBlocks程序和Cygwin64,将源码编译了一下,生成Windows7 环境下可执行的文件webbench.exe ,方便在一个平台下同时使用2个程序,进行网站性能测试。

 

2、下载地址

Windows7 64位下的两个网站性能测试小工具下载地址:

链接:https://pan.baidu.com/s/1hDhBlNc4ss_sp-qPYk_kEQ

提取码:upas

 

3、运行截图:

扫描二维码关注公众号,回复: 3917500 查看本文章

01-Windows7 64位下webbench测试结果

 01-Windows7 64位下webbench测试结果.jpg

02-Apache Bench 性能测试小工具-运行参数

 02-apache bench 性能测试小工具-运行参数.jpg

03-Apache Bench 性能测试小工具-测试结果1

 03-apache bench 性能测试小工具-测试结果1.jpg

04-Apache Bench 性能测试小工具-测试结果2

 04-apache bench 性能测试小工具-测试结果2.jpg

 

4、测试命令及结果分析

ab.exe -n 15000 -c 100 http://www.abc.com/

# -n 15000总共发送的1.5万个访问请求N

# -c 100模拟100个用户同时访问C个;

# 结果分析:

#   Concurrency Level:  模拟并发用户数(命令行中输入的数据)C

#   Time taken for tests: 完成整个测试的时间T

#   Requests per second: 服务器性能(在当前并发下每秒能处理的请求数)=N/T

#   Time per request: (mean)从用户的角度计算,每个请求平均花费时间=T/N/C
#       解释:每个用户发送N/C个请求,共花费T秒,每个请求平均花费时间=T/N/C

#   Time per request:(mean, across all concurrent requests)从服务器角度计算,完成每个请求的平均花费时间=T/N

#   Transfer rate: 传送速率

 

webbench.exe -c 100 -t 5 http://www.abc.com/

# -c 100模拟100个用户同时访问;

# -t 5 持续5秒钟测试时间

# 需要测试的网址 http://www.abc.com/ ,注意只测网站名时,结尾要加/

#

# Speed pages/min 每分钟多少个请求, Speed pages/sec 每秒钟多少个请求

# bytes/sec  传送速率

# Requests: 总共完成的请求数

 

5、注意事项

!!!不要对公共网址发送大量访问请求!!!可用于自己的网站上线前的性能测试。

 

6、使用帮助

C:\>ab --help

ab: wrong number of arguments

Usage: ab [options] [http://]hostname[:port]/path

Options are:

    -n requests     Number of requests to perform

    -c concurrency  Number of multiple requests to make at a time

    -t timelimit    Seconds to max. to spend on benchmarking

                    This implies -n 50000

    -s timeout      Seconds to max. wait for each response

                    Default is 30 seconds

    -b windowsize   Size of TCP send/receive buffer, in bytes

    -B address      Address to bind to when making outgoing connections

    -p postfile     File containing data to POST. Remember also to set -T

    -u putfile      File containing data to PUT. Remember also to set -T

    -T content-type Content-type header to use for POST/PUT data, eg.

                    'application/x-www-form-urlencoded'

                    Default is 'text/plain'

    -v verbosity    How much troubleshooting info to print

    -w              Print out results in HTML tables

    -i              Use HEAD instead of GET

    -x attributes   String to insert as table attributes

    -y attributes   String to insert as tr attributes

    -z attributes   String to insert as td or th attributes

    -C attribute    Add cookie, eg. 'Apache=1234'. (repeatable)

    -H attribute    Add Arbitrary header line, eg. 'Accept-Encoding: gzip'

                    Inserted after all normal header lines. (repeatable)

    -A attribute    Add Basic WWW Authentication, the attributes

                    are a colon separated username and password.

    -P attribute    Add Basic Proxy Authentication, the attributes

                    are a colon separated username and password.

    -X proxy:port   Proxyserver and port number to use

    -V              Print version number and exit

    -k              Use HTTP KeepAlive feature

    -d              Do not show percentiles served table.

    -S              Do not show confidence estimators and warnings.

    -q              Do not show progress when doing more than 150 requests

    -l              Accept variable document length (use this for dynamic pages)

 

    -g filename     Output collected data to gnuplot format file.

    -e filename     Output CSV file with percentages served

    -r              Don't exit on socket receive errors.

    -m method       Method name

    -h              Display usage information (this message)

 

C:\>webbench --help

webbench [option]... URL

  -f|--force               Don't wait for reply from server.

  -r|--reload              Send reload request - Pragma: no-cache.

  -t|--time <sec>          Run benchmark for <sec> seconds. Default 30.

  -p|--proxy <server:port> Use proxy server for request.

  -c|--clients <n>         Run <n> HTTP clients at once. Default one.

  -9|--http09              Use HTTP/0.9 style requests.

  -1|--http10              Use HTTP/1.0 protocol.

  -2|--http11              Use HTTP/1.1 protocol.

  --get                    Use GET request method.

  --head                   Use HEAD request method.

  --options                Use OPTIONS request method.

  --trace                  Use TRACE request method.

  -?|-h|--help             This information.

  -V|--version             Display program version.

 


猜你喜欢

转载自blog.51cto.com/10321203/2312588