sysbench版本信息：http://www.cndba.cn/Marvinn/article/3336

$ sysbench --version
sysbench 1.0.9

CPU相关信息http://www.cndba.cn/Marvinn/article/3336

# 总核数 = 物理CPU个数 X 每颗物理CPU的核数 
# 总逻辑CPU数 = 物理CPU个数 X 每颗物理CPU的核数 X 超线程数

# 查看物理CPU个数
cat /proc/cpuinfo| grep "physical id"| sort| uniq| wc -l

# 查看每个物理CPU中core的个数(即核数)
cat /proc/cpuinfo| grep "cpu cores"| uniq

# 查看逻辑CPU的个数
cat /proc/cpuinfo| grep "processor"| wc -l

# 查看CPU信息（型号）
cat /proc/cpuinfo | grep name | cut -f2 -d: | uniq -c

一般来说，物理CPU个数×每颗核数就应该等于逻辑CPU的个数，如果不相等的话，则表示服务器的CPU支持超线程技术

查看存储相关信息http://www.cndba.cn/Marvinn/article/3336

# lsblk -d -o name,rota
NAME    ROTA
sda        1
sr0        1
nvme0n1    0
nvme1n1    0

rota值如果返回1则表示磁盘可旋转，那么就是HDD了；反之，如果返回0，则表示磁盘不可以旋转，那么就有可能是SSD了

查看具体磁盘的大小
# df -h|grep nvme
/dev/nvme0n1    3.6T   11G  3.4T   1% /data
/dev/nvme1n1    3.6T   11G  3.4T   1% /data2

查看内存相关信息

# free -m
              total        used        free      shared  buff/cache   available
Mem:         193163        7369      154988         474       30805      184382
Swap:             0           0           0

Total表示物理机器内存总大小

网卡相关信息

命令格式:    ethtool 网络接口名
# ethtool em1
Settings for em1:
        Supported ports: [ FIBRE ]
        Supported link modes:   1000baseT/Full 
                                10000baseT/Full 
        Supported pause frame use: Symmetric Receive-only
        Supports auto-negotiation: Yes
        Supported FEC modes: Not reported
        Advertised link modes:  Not reported
        Advertised pause frame use: No
        Advertised auto-negotiation: No
        Advertised FEC modes: Not reported
        Speed: 10000Mb/s
        Duplex: Full
        Port: FIBRE
        PHYAD: 1
        Transceiver: internal
        Auto-negotiation: off
        Supports Wake-on: g
        Wake-on: d
        Current message level: 0x00000000 (0)

        Link detected: yes

Speed: 10000Mb/s 指表示万兆网卡

TiDB集群机器CPU相关信息(CPU配置一样)http://www.cndba.cn/Marvinn/article/3336

# cat /proc/cpuinfo |grep "physical id"|sort |uniq|wc -l  
2
# cat /proc/cpuinfo |grep "cores"|uniq  
cpu cores       : 10
# cat /proc/cpuinfo |grep "processor"|wc -l  
40
# cat /proc/cpuinfo | grep name | cut -f2 -d: | uniq -c
     40  Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz

服务器的CPU支持超线程技术

TiDB集群机器相关信息

TiDB集群部署方案以及配置参数http://www.cndba.cn/Marvinn/article/3336http://www.cndba.cn/Marvinn/article/3336

// TiDB 部署方案
172-16-30-86    1*tidb    1*pd      1*sysbench   1*haproxy
172-16-30-89    1*tidb    2*tikv    1*sysbench  

// 每个物理节点只规划deploy_dir同一个目录，都只有一块盘：
data：NVME SSD （pd、tikv、tidb）

// TiKV 参数配置以及说明
# true (default value) for high reliability, this can prevent data loss when power failure.
sync-log = true

# size of thread pool for grpc server
grpc-concurrency = 4

# The number of connections with each tikv server to send raft messages.
grpc-raft-conn-num = 24

# block-cache used to cache uncompressed blocks, big block-cache can speed up read.in 
# normal cases should tune to 30%-50% system's total memory.
[defaultcf]
block-cache-size = "5GB"

# in normal cases should tune to 10%-30% system's total memory.
[writecf]
block-cache-size: "8000MB"

# should tune to 256MB~2GB.
[raftdb.defaultcf]
block-cache-size: "256MB"

# set store capacity, if no set, use disk capacity.
capacity: 0

readpool:
  storage:
    # size of thread pool for high-priority read operations
    # high-concurrency: 4
    # size of thread pool for normal-priority read operations
    # normal-concurrency: 4
    # size of thread pool for low-priority read operations
    # low-concurrency: 4
  coprocessor:
    # Notice: if CPU_NUM > 8, default thread pool size for coprocessors
    # will be set to CPU_NUM * 0.8.
      high-concurrency: 4
      normal-concurrency: 4
      low-concurrency: 4

sysbench压测结果参数解析http://www.cndba.cn/Marvinn/article/3336http://www.cndba.cn/Marvinn/article/3336http://www.cndba.cn/Marvinn/article/3336

测试结果分析：
SQL statistics:
##queries performed:  #性能统计
read:1994860   #总 select 数量
write: 0       #总update、insert、delete语句数量
other: 284980  #commit、unlock tables以及其他mutex的数量
total: 2279840 #总的执行语句数
transactions:142490 (1186.20 per sec.)  #总的事物数（每秒处理事物数）|通常需要关注的数字(TPS)
queries: 2279840 (18979.13 per sec.)    #读写请求次数（每秒的读写次数）|通常需要关注的数字(QPS)
other operations: 9375 (156.13 per sec.)#其它操作的每秒执行数
ignored errors:0(0.00 per sec.)         #忽略的错误数
reconnects:0(0.00 per sec.)
##General statistics:
total time:120.1216s                    #即time指定的压测总时间
total number of events:142490           #总的事件数,一般与transactions相同
total time taken by event execution: 600.0783s　　 #所有事务耗时相加(不考虑并行因素)
##response time:　　　　//应答时间
min:8.52
avg:107.84
max:480.08
95th percentile:170.48                  #95%的语句的平均响应时间
sum:15365843.76
##Threads fairness:
events (avg/stddev): 1113.2031/14.76
execution time (avg/stddev): 120.0457/0.04

我们一般关注的指标主要有:

response time avg: 平均响应时间。(后面的95%的大小可以通过–percentile=98的方式去更改)
transactions: 精确的说是这一项后面的TPS 。但如果使用了-skip-trx=on,这项事务数恒为0,需要用total number of events 去除以总时间,得到tps(其实还可以分为读tps和写tps)
queries: 用它除以总时间,得到吞吐量QPS

版权声明：本文为博主原创文章，未经博主允许不得转载。