redis-benchmark压力测试

01测试-redis-benchmark压力测试

redis自带有redis-benchmark工具做压力测试，经常用来测试新版本，新特性对基准测试性能的影响。参数场景变化下的性能状况。

主要参数：
 -h <hostname>      服务器地址 (default 127.0.0.1)
 -p <port>          端口 (default 6379)
 -s <socket>        socket (overrides host and port)
 -a <password>      密码
 -c <clients>       并行连接数(default 50)
 -n <requests>      总请求数 (default 100000)
 -d <size>          SET/GET请求value大小，单位：bytes (default 3)
 --dbnum <db>       SELECT 指定库 (default 0)
 -k <boolean>       是否采用keep alive模式，0是，1为reconnect (default 1)
 -r <keyspacelen>   为sadd集合操作SET/GET/INCR随机产生key，随机产生value。
 -P <numreq>        pipeline请求数. Default 1 (no pipeline).
 -e                 输出错误信息，每秒显示不超过1个错误。
 -q                 Quiet. 仅仅查看每秒的查询数。
 --csv              输出为CSV格式。
 -l                 循环次数。
 -t <tests>         指定测试命令。
 -I                 打开空连接，并等待。

Examples:

1.进行20个并发请求，10万请求量，测试set、get、incr、lpush,rpush,lpop、rpop、sadd、hset、
spop、range前100个元素、range前300个元素、range前500个元素、range前600个元素、mset。比较全面的测试
   $ redis-benchmark -h 192.168.1.1 -p 6379 -n 100000 -c 20

2.测试set随机数的性能。
   $ redis-benchmark -t set -n 1000000 -r 100000000

3.测试ping,set,get 10万请求性能，csv输出：
   $ redis-benchmark -t ping,set,get -n 100000 --csv

4.测试指定命令性能:
   $ redis-benchmark -r 10000 -n 10000 eval 'return redis.call("ping")' 0

5.测试list随机10000个元素写入lpush:
   $ redis-benchmark -r 10000 -n 10000 lpush mylist __rand_int__

测试记录：
本次测试版本，配置等。

[root@rcsy ~]# redis-cli -a mysql info | grep version
redis_version:4.0.14
gcc_version:4.8.5

内存2G，2颗核心，SSD：30G。

[root@rcsy redis-4.0.14]# redis-benchmark -n 100000 -c 20 -a mysql
====== PING_INLINE ======
  100000 requests completed in 1.99 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.94% <= 1 milliseconds
99.96% <= 2 milliseconds
100.00% <= 3 milliseconds
50251.26 requests per second

====== PING_BULK ======
  100000 requests completed in 2.13 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.92% <= 1 milliseconds
99.98% <= 2 milliseconds
100.00% <= 2 milliseconds
46970.41 requests per second

====== SET ======
  100000 requests completed in 2.02 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.94% <= 1 milliseconds
99.98% <= 2 milliseconds
99.99% <= 3 milliseconds
100.00% <= 3 milliseconds
49431.54 requests per second

====== GET ======
  100000 requests completed in 2.12 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.91% <= 1 milliseconds
100.00% <= 2 milliseconds
100.00% <= 2 milliseconds
47236.65 requests per second

====== INCR ======
  100000 requests completed in 2.09 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.96% <= 1 milliseconds
99.98% <= 2 milliseconds
100.00% <= 2 milliseconds
47801.15 requests per second

====== LPUSH ======
  100000 requests completed in 2.06 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.87% <= 1 milliseconds
99.96% <= 2 milliseconds
99.99% <= 3 milliseconds
100.00% <= 3 milliseconds
48567.27 requests per second

====== RPUSH ======
  100000 requests completed in 2.05 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.93% <= 1 milliseconds
99.98% <= 2 milliseconds
100.00% <= 2 milliseconds
48685.49 requests per second

====== LPOP ======
  100000 requests completed in 2.09 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.95% <= 1 milliseconds
99.98% <= 2 milliseconds
100.00% <= 2 milliseconds
47915.67 requests per second

====== RPOP ======
  100000 requests completed in 2.11 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.93% <= 1 milliseconds
99.98% <= 2 milliseconds
99.99% <= 4 milliseconds
100.00% <= 4 milliseconds
47438.33 requests per second

====== SADD ======
  100000 requests completed in 2.14 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.90% <= 1 milliseconds
100.00% <= 2 milliseconds
100.00% <= 2 milliseconds
46620.05 requests per second

====== HSET ======
  100000 requests completed in 2.08 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.91% <= 1 milliseconds
99.99% <= 2 milliseconds
100.00% <= 2 milliseconds
47984.64 requests per second

====== SPOP ======
  100000 requests completed in 2.17 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.92% <= 1 milliseconds
99.98% <= 2 milliseconds
100.00% <= 3 milliseconds
100.00% <= 3 milliseconds
46146.75 requests per second

====== LPUSH (needed to benchmark LRANGE) ======
  100000 requests completed in 2.12 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

99.93% <= 1 milliseconds
99.96% <= 2 milliseconds
99.97% <= 3 milliseconds
99.99% <= 4 milliseconds
100.00% <= 4 milliseconds
47281.32 requests per second

====== LRANGE_100 (first 100 elements) ======
  100000 requests completed in 6.34 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

61.85% <= 1 milliseconds
99.77% <= 2 milliseconds
99.98% <= 3 milliseconds
100.00% <= 3 milliseconds
15782.83 requests per second

====== LRANGE_300 (first 300 elements) ======
  100000 requests completed in 15.55 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

1.24% <= 1 milliseconds
39.79% <= 2 milliseconds
92.01% <= 3 milliseconds
98.24% <= 4 milliseconds
99.87% <= 5 milliseconds
99.95% <= 6 milliseconds
99.96% <= 7 milliseconds
99.97% <= 8 milliseconds
99.98% <= 9 milliseconds
99.98% <= 10 milliseconds
99.98% <= 11 milliseconds
99.99% <= 12 milliseconds
99.99% <= 15 milliseconds
99.99% <= 16 milliseconds
99.99% <= 24 milliseconds
100.00% <= 25 milliseconds
100.00% <= 26 milliseconds
6431.28 requests per second

====== LRANGE_500 (first 450 elements) ======
  100000 requests completed in 22.25 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

0.76% <= 1 milliseconds
15.27% <= 2 milliseconds
50.68% <= 3 milliseconds
83.07% <= 4 milliseconds
93.83% <= 5 milliseconds
97.42% <= 6 milliseconds
99.28% <= 7 milliseconds
99.69% <= 8 milliseconds
99.82% <= 9 milliseconds
99.88% <= 10 milliseconds
99.91% <= 11 milliseconds
99.92% <= 12 milliseconds
99.93% <= 13 milliseconds
99.94% <= 14 milliseconds
99.94% <= 15 milliseconds
99.95% <= 16 milliseconds
99.98% <= 17 milliseconds
99.99% <= 18 milliseconds
99.99% <= 19 milliseconds
99.99% <= 30 milliseconds
99.99% <= 31 milliseconds
100.00% <= 32 milliseconds
100.00% <= 32 milliseconds
4494.38 requests per second

====== LRANGE_600 (first 600 elements) ======
  100000 requests completed in 29.40 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

0.04% <= 1 milliseconds
1.65% <= 2 milliseconds
13.74% <= 3 milliseconds
48.50% <= 4 milliseconds
78.79% <= 5 milliseconds
92.05% <= 6 milliseconds
95.91% <= 7 milliseconds
98.48% <= 8 milliseconds
99.74% <= 9 milliseconds
99.86% <= 10 milliseconds
99.89% <= 11 milliseconds
99.90% <= 12 milliseconds
99.91% <= 13 milliseconds
99.92% <= 14 milliseconds
99.93% <= 15 milliseconds
99.94% <= 16 milliseconds
99.96% <= 17 milliseconds
99.97% <= 18 milliseconds
99.98% <= 19 milliseconds
99.98% <= 20 milliseconds
100.00% <= 21 milliseconds
100.00% <= 21 milliseconds
3401.82 requests per second

====== MSET (10 keys) ======
  100000 requests completed in 3.61 seconds
  20 parallel clients
  3 bytes payload
  keep alive: 1

98.40% <= 1 milliseconds
99.93% <= 2 milliseconds
100.00% <= 2 milliseconds
27670.17 requests per second

基本上ping、set、get、lpush、lpop、spop等都达到4-5万多rps，但lrange前100、300、500等就比较慢了，
随着范围越大，月底。从16000--6400---4500--3400，可以看到100-300下降一半多。

压力情况：
   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 24657 root      20   0   16908   5788    968 S  99.3  0.3   0:27.31 redis-benchmark
 24624 root      20   0  149408  10744   1288 R  95.4  0.6   2:11.72 redis-server

 可以看到，CPU基本被占满。 按照正常来讲，至少set，get至少上8万。

[root@rcsy ~]# redis-benchmark -t set -n 1000000 -r 100000000
====== SET ======
  1000000 requests completed in 21.45 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

99.03% <= 1 milliseconds
99.96% <= 2 milliseconds
99.98% <= 3 milliseconds
99.98% <= 4 milliseconds
99.99% <= 5 milliseconds
99.99% <= 8 milliseconds
100.00% <= 9 milliseconds
100.00% <= 10 milliseconds
100.00% <= 12 milliseconds
100.00% <= 12 milliseconds
46620.05 requests per second

[root@rcsy ~]# redis-benchmark -t ping,set,get -n 100000 --csv
"PING_INLINE","48899.75"
"PING_BULK","45351.48"
"SET","44943.82"
"GET","45351.48"

[root@rcsy ~]# redis-benchmark -r 10000 -n 10000 eval 'return redis.call("ping")' 0
====== eval return redis.call("ping") 0 ======
  10000 requests completed in 0.22 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

98.29% <= 1 milliseconds
99.51% <= 3 milliseconds
99.74% <= 4 milliseconds
99.96% <= 5 milliseconds
100.00% <= 5 milliseconds
46082.95 requests per second

[root@rcsy ~]# redis-benchmark -r 10000 -n 10000 lpush mylist __rand_int__
====== lpush mylist __rand_int__ ======
  10000 requests completed in 0.22 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

98.80% <= 1 milliseconds
99.51% <= 2 milliseconds
99.80% <= 3 milliseconds
100.00% <= 3 milliseconds
46082.95 requests per second