虚机性能测试

安装sysbench-1.0.4

下载URL：https://github.com/akopytov/sysbench
安装sysbench之前需要安装以下依赖包：m4 autoconf automake libtool 。
使用 yum install 安装这几个依赖包。

# tar zxvf sysbench-1.0.4.tar.gz
# cd sysbench-1.0.4
# ./autogen.sh 
# ./configure --without-mysql          //不需要测试数据库，所以用--without选项
# make
# make install

虚拟机环境

• IP：10.37.64.53

• CUP：8 core

• Mem：18G

• OS：RHEL 6.3 (64 Bit)

  -bash-4.1$ cat /proc/cpuinfo

  processor : 0
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 1
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 2
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 3
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 4
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 5
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 6
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

  processor : 7
  vendor_id : GenuineIntel
  cpu family : 6
  model : 42
  model name : Intel Xeon E312xx (Sandy Bridge)
  stepping : 1
  cpu MHz : 2394.454
  cache size : 4096 KB
  fpu : yes
  fpu_exception : yes
  cpuid level : 13
  wp : yes
  flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb lm constant_tsc rep_good unfair_spinlock pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm xsaveopt fsgsbase bmi1 smep erms
  bogomips : 4788.90
  clflush size : 64
  cache_alignment : 64
  address sizes : 46 bits physical, 48 bits virtual
  power management:

CPU性能

利用sysbench评测CPU性能，线程数设成和CPU核数相同为8

-bash-4.1$ sysbench --threads=8 --events=10000 --time=120 --test=cpu --cpu-max-prime=200000 run
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any 
options.
sysbench 1.0.4 (using bundled LuaJIT 2.1.0-beta2)

Running the test with following options:
Number of threads: 8
Initializing random number generator from current time


Prime numbers limit: 200000

Initializing worker threads...

Threads started!


General statistics:
total time:                          116.4856s
total number of events:              10000

Latency (ms):
     min:                                 92.54
     avg:                                 93.15
     max:                                119.90
     95th percentile:                     94.10
     sum:                             931541.67

Threads fairness:
	events (avg/stddev):           1250.0000/0.71
	execution time (avg/stddev):   116.4427/0.03

内存性能

-bash-4.1$ sysbench --threads=8 --test=memory --memory-block-size=4K --memory-total-size=10G --memory-

oper=write run
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any	options.
sysbench 1.0.4 (using bundled LuaJIT 2.1.0-beta2)

Running the test with following options:
Number of threads: 8
Initializing random number generator from current time

Initializing worker threads...

Threads started!

Operations performed: 2621440 (2111415.90 ops/sec)

10240.00 MiB transferred (8247.72 MiB/sec)


General statistics:
	total time:                          1.2393s
	total number of events:              2621440

Latency (ms):
     min:                                  0.00
     avg:                                  0.00
     max:                                  0.85
     95th percentile:                      0.00
     sum:                               4325.52

Threads fairness:
	events (avg/stddev):           327680.0000/61070.57
	execution time (avg/stddev):   0.5407/0.10

通过设置memory-block-size的值（如8K、4K、1K、128),测试一下内存写性能。

存储性能

pg_test_fsync

使用PostgreSQL自带的工具pg_test_fsync测试前，需要cd到需要测试的文件系统的某个目录下。

-bash-4.1$ pg_test_fsync
5 seconds per test
O_DIRECT supported on this platform for open_datasync and open_sync.

Compare file sync methods using one 8kB write:
(in wal_sync_method preference order, except fdatasync is Linux's default)
    open_datasync                      7495.343 ops/sec     133 usecs/op
    fdatasync                          7159.733 ops/sec     140 usecs/op
    fsync                              7675.694 ops/sec     130 usecs/op
    fsync_writethrough                            n/a
    open_sync                          8455.910 ops/sec     118 usecs/op

Compare file sync methods using two 8kB writes:
(in wal_sync_method preference order, except fdatasync is Linux's default)
    open_datasync                      3994.309 ops/sec     250 usecs/op
    fdatasync                          7287.177 ops/sec     137 usecs/op
    fsync                              6976.354 ops/sec     143 usecs/op
    fsync_writethrough                            n/a
    open_sync                          3936.843 ops/sec     254 usecs/op

Compare open_sync with different write sizes:
(This is designed to compare the cost of writing 16kB in different write
open_sync sizes.)
     1 * 16kB open_sync write          7511.188 ops/sec     133 usecs/op
     2 *  8kB open_sync writes         3880.634 ops/sec     258 usecs/op
     4 *  4kB open_sync writes         1954.984 ops/sec     512 usecs/op
     8 *  2kB open_sync writes          965.620 ops/sec    1036 usecs/op
    16 *  1kB open_sync writes          489.129 ops/sec    2044 usecs/op

Test if fsync on non-write file descriptor is honored:
(If the times are similar, fsync() can sync data written on a different
descriptor.)
    write, fsync, close                7481.958 ops/sec     134 usecs/op
    write, close, fsync                7480.560 ops/sec     134 usecs/op

Non-sync'ed 8kB writes:
    write                            370970.939 ops/sec       3 usecs/op

sysbench fileio

生成数据

由于会读到缓存数据，影响测试结果。所以使用events来控制读请求次数。本次设为100000，由于block-size为8K，所以读取的数据总量为800M左右。

[root@sndsdevsdd02 test]# sysbench --threads=10 --events=100000 --test=fileio --file-block-size=8192 --file-test-mode=seqrd --file-fsync-mode=fdatasync prepare

清理缓存

读之前需要清理缓存，否则测试数据没意义

[root@sndsdevsdd02 test]# echo 1 > /proc/sys/vm/drop_caches

测试

[root@sndsdevsdd02 test]# sysbench --threads=10 --events=100000 --test=fileio --file-block-size=8192 --file-test-mode=seqrd --file-fsync-mode=fdatasync run

清理数据

[root@sndsdevsdd02 test]# sysbench --threads=10 --events=100000 --test=fileio --file-block-size=8192 --file-test-mode=seqrd --file-fsync-mode=fdatasync cleanup

修改参数，多次测试

threads=1，10，100，1000
file-block-size=8192
file-test-mode=seqwr,seqrd,rndwr,rndrd
file-fsync-mode=fdatasync

网络性能

安装netperf

下载URL：ftp://ftp.netperf.org/netperf/

安装nmon

下载URL：http://nmon.sourceforge.net/pmwiki.php?n=Site.Download

启动服务端

在10.37.64.54机器上启动，使用默认端口

netserver -D

传输速度

传输速度为770Mbit/s

[root@sndsdevsdd02 ~]# netperf -t TCP_STREAM -H 10.37.64.54 -l 5 -Cc
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.37.64.54 () port 0 AF_INET
Recv   Send    Send                          Utilization       Service Demand
Socket Socket  Message  Elapsed              Send     Recv     Send    Recv
Size   Size    Size     Time     Throughput  local    remote   local   remote
bytes  bytes   bytes    secs.    10^6bits/s  % S      % S      us/KB   us/KB

87380  16384  16384    5.00        769.73   0.31     1.50     0.262   1.278

请求响应速度

单个连接

[root@sndsdevsdd02 ~]# netperf -t TCP_RR -H 10.37.64.54 -l 5 -Cc
MIGRATED TCP REQUEST/RESPONSE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.37.64.54 () port 0 AF_INET : 

first burst 0
Local /Remote
Socket Size   Request Resp.  Elapsed Trans.   CPU    CPU    S.dem   S.dem
Send   Recv   Size    Size   Time    Rate     local  remote local   remote
bytes  bytes  bytes   bytes  secs.   per sec  % S    % S    us/Tr   us/Tr

16384  87380  1       1      10.00   9862.43  1.23   1.29   9.945   10.430
16384  87380

多个连接的时候(同一个网段，连接时间设为10s)，通过nmon查看网络packin/packout的最大值

连接数		pkgs/s
1			 10087
10			 74444
30			108130
60			165176
100			206103
150			221118
200			228998

连接数为200时，其中一个连接的输出如下：

MIGRATED TCP REQUEST/RESPONSE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.37.64.54 () port 0 AF_INET : 

first burst 0
Local /Remote
Socket Size   Request Resp.  Elapsed Trans.   CPU    CPU    S.dem   S.dem
Send   Recv   Size    Size   Time    Rate     local  remote local   remote
bytes  bytes  bytes   bytes  secs.   per sec  % S    % S    us/Tr   us/Tr

16384  87380  1       1      10.00   1015.97  32.63  32.67  2569.163 2572.770
16384  87380