sysbench工具手把手安装配置及使用…

sysbench是一款测试工具

主要包括以下几种方式的测试：

1、cpu性能

2、磁盘io性能

3、调度程式性能

4、内存分配及传输速度

5、POSIX线程性能

6、数据库性能(OLTP基准测试)现在sysbench主要支持 MySQL,pgsql,oracle 这3种数据库

本文以实际的安装和简单的测试以做抛砖引玉，希望在泰岳的性能测试中发挥应有的作用。

 

一：安装mysql

先安装

MySQL-client-5.5.32-1.rhel5.i386.rpm

MySQL-server-5.5.32-1.rhel5.i386.rpm

下面三个用于性能测试用

MySQL-shared-5.5.32-1.rhel5.i386.rpm

MySQL-devel-5.5.32-1.rhel5.i386.rpm

MySQL-shared-compat-5.5.32-1.rhel5.i386.rpm

二：安装sysbench

   
安装sysbench因为是源码需要编译及需要其他的包或工具，较为繁琐，现做具体介绍

1.下载sysbench 
sysbench-0.4.12.tar.gz，并上传到/usr/local目录下

[root@localhost local]# ls

bin                   
etc      
man                                         
sbin

cmake-2.8.10.2        
games    
mysql-5.6.11.tar.gz                         
share

cmake-2.8.10.2.tar.gz 
include  
MySQL-client-5.5.32-1.rhel5.i386.rpm        
src

config                
lib      
MySQL-devel-5.5.32-1.rhel5.i386.rpm         

config.log            
libexec  
MySQL-server-5.5.32-1.rhel5.i386.rpm        
sysbench-0.4.12.tar.gz

config.status         
libtool  
MySQL-shared-5.5.32-1.rhel5.i386.rpm

doc                   
Makefile 
MySQL-shared-compat-5.5.32-1.rhel5.i386.rpm

 

2.解压   tax -zxvf
sysbench-0.4.12.tar.gz

[root@localhost local]# ls

bin                   
etc      
man                                         
sbin

cmake-2.8.10.2        
games    
mysql-5.6.11.tar.gz                         
share

cmake-2.8.10.2.tar.gz 
include  
MySQL-client-5.5.32-1.rhel5.i386.rpm        
src

config                
lib      
MySQL-devel-5.5.32-1.rhel5.i386.rpm         
sysbench-0.4.12

config.log            
libexec  
MySQL-server-5.5.32-1.rhel5.i386.rpm        
sysbench-0.4.12.tar.gz

config.status         
libtool  
MySQL-shared-5.5.32-1.rhel5.i386.rpm

doc                   
Makefile 
MySQL-shared-compat-5.5.32-1.rhel5.i386.rpm

3.编译，安装

/usr/local/sysbench-0.4.12/configure
--with-mysql-includes=/usr/include/mysql --with-mysql-libs=/usr/lib
&& make
&& make install

....

test -z "/usr/local/share/doc/sysbench" || mkdir -p --
"/usr/local/share/doc/sysbench"

 /usr/bin/install -c -m 644 'manual.html'
'/usr/local/share/doc/sysbench/manual.html'

make[3]: Leaving directory `/usr/local/sysbench-0.4.12/doc'

make[2]: Leaving directory `/usr/local/sysbench-0.4.12/doc'

make[1]: Leaving directory `/usr/local/sysbench-0.4.12/doc'

Making install in sysbench

make[1]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench'

Making install in drivers

make[2]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/drivers'

Making install in mysql

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/drivers/mysql'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/drivers/mysql'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/drivers/mysql'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/drivers/mysql'

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/drivers'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/drivers'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/drivers'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/drivers'

make[2]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/drivers'

Making install in tests

make[2]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests'

Making install in cpu

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/cpu'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/cpu'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/cpu'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/cpu'

Making install in fileio

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/fileio'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/fileio'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/fileio'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/fileio'

Making install in memory

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/memory'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/memory'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/memory'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/memory'

Making install in threads

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/threads'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/threads'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/threads'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/threads'

Making install in mutex

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/mutex'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/mutex'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/mutex'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/mutex'

Making install in oltp

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/oltp'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests/oltp'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/oltp'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests/oltp'

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests'

make[4]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench/tests'

make[4]: Nothing to be done for `install-exec-am'.

make[4]: Nothing to be done for `install-data-am'.

make[4]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests'

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests'

make[2]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench/tests'

Making install in .

make[2]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench'

make[3]: Entering directory
`/usr/local/sysbench-0.4.12/sysbench'

test -z "/usr/local/bin" || mkdir -p -- "/usr/local/bin"

  /bin/sh ../libtool --mode=install
/usr/bin/install -c 'sysbench' '/usr/local/bin/sysbench'

/usr/bin/install -c sysbench /usr/local/bin/sysbench

make[3]: Nothing to be done for `install-data-am'.

make[3]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench'

make[2]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench'

make[1]: Leaving directory
`/usr/local/sysbench-0.4.12/sysbench'

make[1]: Entering directory `/usr/local/sysbench-0.4.12'

make[2]: Entering directory `/usr/local/sysbench-0.4.12'

make[2]: Nothing to be done for `install-exec-am'.

make[2]: Nothing to be done for `install-data-am'.

make[2]: Leaving directory `/usr/local/sysbench-0.4.12'

make[1]: Leaving directory `/usr/local/sysbench-0.4.12'

三测试：

1.数据库oltp测试，这个需要创建数据库sbtest;以下见具体测试用例

[root@localhost local]# sysbench --num-threads=5
--max-requests=1000 --test=oltp --mysql-table-engine=myisam
--oltp-table-size=4000000 --mysql-socket=/var/lib/mysql/mysql.sock
--mysql-db=sbtest --mysql-user=root
--mysql-host=127.0.0.1  --mysql-password=root
run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

No DB drivers specified, using mysql

Running the test with following options:

Number of threads: 5

Doing OLTP test.

Running mixed OLTP test

Using Special distribution (12 iterations,  1 pct
of values are returned in 75 pct cases)

Using "LOCK TABLES WRITE" for starting transactions

Using auto_inc on the id column

Maximum number of requests for OLTP test is limited to 1000

Threads started!

Done.

OLTP test statistics:

    queries
performed:

       
read:                           
14000

       
write:                          
5000

       
other:                          
2000

       
total:                          
21000

   
transactions:                       
1000   (51.63 per sec.)

   
deadlocks:                          
0     
(0.00 per sec.)

    read/write
requests:                
19000  (981.00 per sec.)

    other
operations:                   
2000   (103.26 per sec.)

Test execution summary:

    total
time:                         
19.3679s

    total number
of
events:             
1000

    total time
taken by event execution: 94.0158

    per-request
statistics:

        
min:                                
14.71ms

        
avg:                                
94.02ms

        
max:                              
5708.77ms

        
approx.  95
percentile:            
113.91ms

Threads fairness:

    events
(avg/stddev):          
200.0000/72.46

    execution
time (avg/stddev):  
18.8032/0.26

[root@localhost local]# sysbench --num-threads=5
--max-requests=5000 --test=oltp --mysql-table-engine=myisam
--oltp-table-size=4000000 --mysql-socket=/var/lib/mysql/mysql.sock
--mysql-db=sbtest --mysql-user=root
--mysql-host=127.0.0.1  --mysql-password=root
run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

No DB drivers specified, using mysql

Running the test with following options:

Number of threads: 5

Doing OLTP test.

Running mixed OLTP test

Using Special distribution (12 iterations,  1 pct
of values are returned in 75 pct cases)

Using "LOCK TABLES WRITE" for starting transactions

Using auto_inc on the id column

Maximum number of requests for OLTP test is limited to 5000

Threads started!

Done.

OLTP test statistics:

    queries
performed:

       
read:                           
70000

       
write:                          
25000

       
other:                          
10000

       
total:                          
105000

   
transactions:                       
5000   (46.05 per sec.)

   
deadlocks:                          
0     
(0.00 per sec.)

    read/write
requests:                
95000  (874.91 per sec.)

    other
operations:                   
10000  (92.10 per sec.)

Test execution summary:

    total
time:                         
108.5830s

    total number
of
events:             
5000

    total time
taken by event execution: 542.5105

    per-request
statistics:

        
min:                                
17.93ms

        
avg:                               
108.50ms

        
max:                              
7212.44ms

        
approx.  95
percentile:            
146.43ms

Threads fairness:

    events
(avg/stddev):          
1000.0000/96.35

    execution
time (avg/stddev):  
108.5021/0.03

2.cpu测试

cpu测试主要是进行素数的加法运算，在上面的例子中，指定了最大的素数为
20000，自己可以根据机器cpu的性能来适当调整数值。

sysbench --test=cpu --cpu-max-prime=200 run

[root@localhost ~]# sysbench --test=cpu --cpu-max-prime=200
run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Running the test with following options:

Number of threads: 1

Doing CPU performance benchmark

Threads started!

Done.

Maximum prime number checked in CPU test: 200

Test execution summary:

    total
time:                         
0.5087s

    total number
of
events:             
10000

    total time
taken by event execution: 0.4448

    per-request
statistics:

        
min:                                 
0.03ms

        
avg:                                 
0.04ms

        
max:                                
20.45ms

        
approx.  95
percentile:              
0.02ms

Threads fairness:

    events
(avg/stddev):          
10000.0000/0.00

    execution
time (avg/stddev):  
0.4448/0.00

[root@localhost ~]# sysbench --test=cpu --cpu-max-prime=2000
run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Running the test with following options:

Number of threads: 1

Doing CPU performance benchmark

Threads started!

Done.

Maximum prime number checked in CPU test: 2000

Test execution summary:

    total
time:                         
3.3867s

    total number
of
events:             
10000

    total time
taken by event execution: 3.3185

    per-request
statistics:

        
min:                                 
0.14ms

        
avg:                                 
0.33ms

        
max:                                
14.34ms

        
approx.  95
percentile:              
0.36ms

Threads fairness:

    events
(avg/stddev):          
10000.0000/0.00

    execution
time (avg/stddev):  
3.3185/0.00

[root@localhost ~]# sysbench --test=cpu --cpu-max-prime=20000
run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Running the test with following options:

Number of threads: 1

Doing CPU performance benchmark

Threads started!

Done.

Maximum prime number checked in CPU test: 20000

Test execution summary:

    total
time:                         
61.4178s

    total number
of
events:             
10000

    total time
taken by event execution: 61.3427

    per-request
statistics:

        
min:                                 
1.13ms

        
avg:                                 
6.13ms

        
max:                                
56.49ms

        
approx.  95
percentile:              
7.96ms

Threads fairness:

    events
(avg/stddev):          
10000.0000/0.00

    execution
time (avg/stddev):  
61.3427/0.00

3、线程测试

sysbench --test=threads --num-threads=16 --thread-yields=100
--thread-locks=2 run

[root@localhost ~]# sysbench --test=cpu --cpu-max-prime=20000
run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Running the test with following options:

Number of threads: 1

Doing CPU performance benchmark

Threads started!

Done.

Maximum prime number checked in CPU test: 20000

Test execution summary:

    total
time:                         
61.4178s

    total number
of
events:             
10000

    total time
taken by event execution: 61.3427

    per-request
statistics:

        
min:                                 
1.13ms

        
avg:                                 
6.13ms

        
max:                                
56.49ms

        
approx.  95
percentile:              
7.96ms

Threads fairness:

    events
(avg/stddev):          
10000.0000/0.00

    execution
time (avg/stddev):  
61.3427/0.00

[root@localhost ~]# sysbench --test=threads --num-threads=16
--thread-yields=100 --thread-locks=2 run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Running the test with following options:

Number of threads: 16

Doing thread subsystem performance test

Thread yields per test: 100 Locks used: 2

Threads started!

Done.

Test execution summary:

    total
time:                         
83.0173s

    total number
of
events:             
10000

    total time
taken by event execution: 1327.3256

    per-request
statistics:

        
min:                                
13.21ms

        
avg:                               
132.73ms

        
max:                              
7176.51ms

        
approx.  95
percentile:            
835.83ms

Threads fairness:

    events
(avg/stddev):          
625.0000/109.40

    execution
time (avg/stddev):  
82.9578/0.03

[root@localhost ~]#

4、磁盘IO性能测试

sysbench --test=fileio --num-threads=5 --file-total-size=100M
--file-test-mode=rndrw prepare

sysbench --test=fileio --num-threads=5 --file-total-size=100M
--file-test-mode=rndrw run

sysbench --test=fileio --num-threads=5 --file-total-size=100M
--file-test-mode=rndrw  cleanup

[root@localhost ~]# sysbench --test=fileio --num-threads=5
--file-total-size=100M --file-test-mode=rndrw run

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Running the test with following options:

Number of threads: 5

Extra file open flags: 0

128 files, 800Kb each

100Mb total file size

Block size 16Kb

Number of random requests for random IO: 10000

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

Threads started!

Done.

Operations performed:  5999 Read, 4001 Write,
12800 Other = 22800 Total

Read 93.734Mb  Written 62.516Mb 
Total transferred 156.25Mb  (12.456Mb/sec)

  797.20 Requests/sec executed

Test execution summary:

    total
time:                         
12.5439s

    total number
of
events:             
10000

    total time
taken by event execution: 4.0691

    per-request
statistics:

        
min:                                 
0.03ms

        
avg:                                 
0.41ms

        
max:                               
207.37ms

        
approx.  95
percentile:              
0.08ms

Threads fairness:

    events
(avg/stddev):          
2000.0000/247.35

    execution
time (avg/stddev):  
0.8138/0.22

root@localhost ~]# sysbench
--test=fileio --num-threads=5 --file-total-size=100M
--file-test-mode=rndrw  cleanup

sysbench 0.4.12:  multi-threaded system evaluation
benchmark

Removing test files...

[root@localhost ~]#

上述参数指定了最大创建5个线程，创建的文件总大小为100M，文件读写模式为随机读。

4、内存测试

sysbench --test=memory --memory-block-size=8k
--memory-total-size=100M run

上述参数指定了本次测试整个过程是在内存中传输 100M 的数据量，每个 block 大小为 8K。

四：实际过程中可能会遇到一些问题：

问题2. 出现configure: error: no acceptable C compiler found in
$PATH

解决办法：安装gcc,命令为：yum install gcc。

问题2：出现../libtool: line 5172: : command not found

解决办法   运行[root@localhost
sysbench-0.4.12]# ./autogen.sh

问题3：source directory already configured, make distclean的错误解决
.

解决办法：在sysbench目录运行 make distclean

问题4：configure时候出现 i686-pc-linux' not
recognized提示               

     
在运行./configure 时如果出现下面的提示信息：

     
checking build system

     
type... Invalid configuration `i686-pc-linux-': machine
`i686-pc-linux' not

     
recognized

     
configure: error:

     
/bin/sh ./config.sub i686-pc-linux- failed

解决办法：就使用./configure --build=i686-pc-linux-gnu来设置