（十二）高并发redis学习笔记：对redis主从架构redis-benchmark压测

最新推荐文章于 2025-12-03 18:51:31 发布

原创最新推荐文章于 2025-12-03 18:51:31 发布 · 592 阅读

1 ·

本内容遵循CC 4.0 BY-SA版权协议

标签

#redis #分布式 #java #缓存

redis 专栏收录该内容

31 篇文章

订阅专栏

本文介绍了在1核1G的环境下，使用redis-benchmark对一主一从的Redis架构进行压测，探讨了单实例的读写QPS，并讨论了通过增加从节点来提升读取吞吐量的方法。根据测试，单个从节点可处理约5万QPS，双从节点可达到10万+。这是作者的学习笔记，旨在分享个人技术积累，欢迎指正。

前提
1.上一小结，搭建好了一主一从的redis架构
2.redis自己提供的redis-benchmark压测工具，是最快捷最方便的。

1、对redis读写分离架构进行压测，单实例写QPS+单实例读QPS

cd /usr/local/redis-3.2.8/src
./redis-benchmark -h 192.168.8.187

主要的命令格式：
-c 模拟多少个客户端发送请求，默认50
-n 请求总数(默认 100000)
-d 数据大小：(默认 2)

我的机器时1核1G，虚拟机

[root@cache01 redis-3.2.8]# cd /usr/local/redis-3.2.8/src/
[root@cache01 src]# ./redis-benchmark -h 192.168.8.181
====== PING_INLINE ======
  100000 requests completed in 2.55 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

78.09% <= 1 milliseconds
96.26% <= 2 milliseconds
99.45% <= 3 milliseconds
99.80% <= 4 milliseconds
99.88% <= 5 milliseconds
99.90% <= 7 milliseconds
99.95% <= 11 milliseconds
99.96% <= 12 milliseconds
99.99% <= 13 milliseconds
100.00% <= 13 milliseconds
39231.07 requests per second

====== PING_BULK ======
  100000 requests completed in 2.58 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

76.36% <= 1 milliseconds
96.60% <= 2 milliseconds
99.50% <= 3 milliseconds
99.77% <= 4 milliseconds
99.87% <= 5 milliseconds
99.89% <= 6 milliseconds
99.90% <= 8 milliseconds
99.95% <= 9 milliseconds
99.95% <= 13 milliseconds
99.99% <= 14 milliseconds
100.00% <= 14 milliseconds
38759.69 requests per second

====== SET ======
  100000 requests completed in 2.97 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

66.38% <= 1 milliseconds
92.05% <= 2 milliseconds
97.89% <= 3 milliseconds
99.28% <= 4 milliseconds
99.83% <= 5 milliseconds
99.94% <= 6 milliseconds
99.99% <= 7 milliseconds
100.00% <= 7 milliseconds
33681.38 requests per second

====== GET ======
  100000 requests completed in 2.54 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

77.38% <= 1 milliseconds
96.29% <= 2 milliseconds
99.69% <= 3 milliseconds
99.96% <= 4 milliseconds
100.00% <= 5 milliseconds
39401.10 requests per second

====== INCR ======
  100000 requests completed in 2.75 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

72.91% <= 1 milliseconds
95.20% <= 2 milliseconds
98.89% <= 3 milliseconds
99.54% <= 4 milliseconds
99.75% <= 5 milliseconds
99.86% <= 6 milliseconds
99.93% <= 7 milliseconds
99.96% <= 8 milliseconds
99.97% <= 11 milliseconds
99.98% <= 12 milliseconds
99.98% <= 18 milliseconds
99.98% <= 23 milliseconds
99.99% <= 24 milliseconds
100.00% <= 24 milliseconds
36416.61 requests per second

====== LPUSH ======
  100000 requests completed in 2.70 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

73.04% <= 1 milliseconds
95.10% <= 2 milliseconds
98.89% <= 3 milliseconds
99.54% <= 4 milliseconds
99.91% <= 5 milliseconds
99.95% <= 10 milliseconds
99.96% <= 11 milliseconds
99.98% <= 12 milliseconds
100.00% <= 12 milliseconds
36995.93 requests per second

====== RPUSH ======
  100000 requests completed in 3.16 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

59.93% <= 1 milliseconds
90.73% <= 2 milliseconds
97.81% <= 3 milliseconds
99.49% <= 4 milliseconds
99.81% <= 5 milliseconds
99.90% <= 6 milliseconds
99.96% <= 7 milliseconds
99.99% <= 8 milliseconds
100.00% <= 9 milliseconds
31645.57 requests per second

====== LPOP ======
  100000 requests completed in 2.74 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

72.06% <= 1 milliseconds
94.60% <= 2 milliseconds
98.81% <= 3 milliseconds
99.58% <= 4 milliseconds
99.79% <= 5 milliseconds
99.91% <= 6 milliseconds
99.96% <= 7 milliseconds
99.98% <= 8 milliseconds
100.00% <= 8 milliseconds
36469.73 requests per second

====== RPOP ======
  100000 requests completed in 2.70 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

72.58% <= 1 milliseconds
95.13% <= 2 milliseconds
99.14% <= 3 milliseconds
99.74% <= 4 milliseconds
99.85% <= 5 milliseconds
100.00% <= 5 milliseconds
36982.25 requests per second

====== SADD ======
  100000 requests completed in 2.99 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

65.65% <= 1 milliseconds
93.31% <= 2 milliseconds
98.32% <= 3 milliseconds
99.37% <= 4 milliseconds
99.59% <= 5 milliseconds
99.75% <= 6 milliseconds
99.86% <= 7 milliseconds
99.91% <= 8 milliseconds
99.93% <= 9 milliseconds
99.96% <= 12 milliseconds
99.96% <= 13 milliseconds
99.97% <= 14 milliseconds
99.98% <= 51 milliseconds
100.00% <= 51 milliseconds
33411.29 requests per second

====== SPOP ======
  100000 requests completed in 2.81 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

68.93% <= 1 milliseconds
94.58% <= 2 milliseconds
99.08% <= 3 milliseconds
99.77% <= 4 milliseconds
99.93% <= 5 milliseconds
99.98% <= 6 milliseconds
100.00% <= 7 milliseconds
35637.92 requests per second

====== LPUSH (needed to benchmark LRANGE) ======
  100000 requests completed in 3.34 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

54.32% <= 1 milliseconds
89.35% <= 2 milliseconds
97.53% <= 3 milliseconds
99.10% <= 4 milliseconds
99.73% <= 5 milliseconds
99.83% <= 6 milliseconds
99.93% <= 7 milliseconds
99.99% <= 8 milliseconds
100.00% <= 8 milliseconds
29931.16 requests per second

====== LRANGE_100 (first 100 elements) ======
  100000 requests completed in 2.84 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

68.46% <= 1 milliseconds
93.82% <= 2 milliseconds
98.71% <= 3 milliseconds
99.51% <= 4 milliseconds
99.86% <= 5 milliseconds
99.95% <= 6 milliseconds
99.97% <= 7 milliseconds
100.00% <= 7 milliseconds
35174.11 requests per second

====== LRANGE_300 (first 300 elements) ======
  100000 requests completed in 3.09 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

62.49% <= 1 milliseconds
91.04% <= 2 milliseconds
97.55% <= 3 milliseconds
99.15% <= 4 milliseconds
99.74% <= 5 milliseconds
99.92% <= 6 milliseconds
99.95% <= 7 milliseconds
99.96% <= 8 milliseconds
99.99% <= 9 milliseconds
100.00% <= 9 milliseconds
32383.42 requests per second

====== LRANGE_500 (first 450 elements) ======
  100000 requests completed in 3.02 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

66.40% <= 1 milliseconds
92.29% <= 2 milliseconds
97.66% <= 3 milliseconds
99.08% <= 4 milliseconds
99.57% <= 5 milliseconds
99.69% <= 6 milliseconds
99.77% <= 7 milliseconds
99.81% <= 8 milliseconds
99.84% <= 9 milliseconds
99.87% <= 10 milliseconds
99.92% <= 11 milliseconds
99.92% <= 15 milliseconds
99.96% <= 16 milliseconds
99.97% <= 18 milliseconds
99.98% <= 19 milliseconds
99.99% <= 35 milliseconds
100.00% <= 35 milliseconds
33090.67 requests per second

====== LRANGE_600 (first 600 elements) ======
  100000 requests completed in 3.25 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

57.99% <= 1 milliseconds
89.38% <= 2 milliseconds
97.58% <= 3 milliseconds
99.35% <= 4 milliseconds
99.74% <= 5 milliseconds
99.94% <= 6 milliseconds
99.99% <= 7 milliseconds
99.99% <= 8 milliseconds
100.00% <= 9 milliseconds
100.00% <= 9 milliseconds
30788.18 requests per second

====== MSET (10 keys) ======
  100000 requests completed in 2.97 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

59.27% <= 1 milliseconds
92.98% <= 2 milliseconds
98.76% <= 3 milliseconds
99.79% <= 4 milliseconds
99.95% <= 5 milliseconds
99.95% <= 7 milliseconds
99.96% <= 8 milliseconds
99.98% <= 9 milliseconds
100.00% <= 9 milliseconds
33704.08 requests per second

大概可以看出，各种请求在某个时间阈值内返回的数据，大概占比多少。以及每秒钟的请求数量多少。

2、水平扩容redis读节点，提升度吞吐量
一般来说，单个从节点读请QPS在5万左右，两个redis从节点，所有的读请求打到两台机器上去，承载整个集群读QPS在10万+。
此文章仅代表自己（本菜鸟）学习积累记录，或者学习笔记，如有侵权，请联系作者删除。人无完人，文章也一样，文笔稚嫩，在下不才，勿喷，如果有错误之处，还望指出，感激不尽~

技术之路不在一时，山高水长，纵使缓慢，驰而不息。

公众号：秦怀杂货店