本文介绍系统缓存组件,采用NOSQL之Redis作为系统缓存层。
一、背景
系统考虑到高并发的使用场景。对于并发提交场景,通过上一章节介绍的RabbitMQ组件解决。对于系统高并发查询,为了提供性能减少数据库压力,我们加入缓存机制,可以不同层次加入缓存支持,本文主要介绍应用服务层和数据层之间加入缓存机制提升性能。业界缓存组件有Redis、Memcached、MemoryCache。本系统采用Redis缓存组件,有些系统将Redis当作MQ使用,此场景本系统用RabbitMQ,Redis主要用于系统缓存应用。
二、Redis简介
Redis是一个开源的Key-Value数据库,使用C语言编写、支持网络、可基于内存亦可持久化的NOSQL数据库,并提供多种语言的API,例如:Java,C/C++,C#,PHP,JavaScript,Perl,Object-C,Python,Ruby等语言驱动。自Redis3.0开始支持集群方案。
相关Redis Cluster 原理在此不多介绍,网络上有很多资料。
三、Redis集群应用
(一)环境介绍
本系统基于Linux之CentOS搭建Redis3.0集群。将三个Instance部署于一台虚拟机,应用部署于windows平台。
序号 | 服务IP | 说明 |
1 | 192.168.1.110 | Redis节点A端口:7000(M),7003(S) Redis节点B端口:7001(M),7004(S) Redis节点C端口:7002(M),7005(S) |
(二)安装Redis
1、安装相关依赖工具
[root@andoncentos 桌面]# yum -y install gcc openssl-devel libyaml-devel libffi-devel readline-devel zlib-devel gdbm-devel ncurses-devel gcc-c++ automake autoconf
2、安装Redis 3.0.6
[root@andoncentos 桌面]# cd /usr/loacal
[root@andoncentos loacal]# wget http://download.redis.io/releases/redis-3.0.6.tar.gz
[root@andoncentos 桌面]# tar xvf redis-3.0.6.tar.gz
[root@andoncentos 桌面]# cd redis-3.0.6/
[root@andoncentos 桌面]# make MALLOC=libc
[root@andoncentos redis-3.0.6]# make install
3、由于我们使用不同端口号区分,在两个服务各自建立以端口命名的文件夹。配置7000节点服务,将redis-server和redis.conf复制到/etc/redis/7000
[root@andoncentos 桌面]# mkdir /etc/redis/7000
[root@andoncentos 桌面]# mkdir /etc/redis/7001
[root@andoncentos 桌面]# mkdir /etc/redis/7002
[root@andoncentos 桌面]# mkdir /etc/redis/7003
[root@andoncentos 桌面]# mkdir /etc/redis/7004
[root@andoncentos 桌面]# mkdir /etc/redis/7005
[root@andoncentos redis-3.0.6]# cp /usr/local/redis-3.0.6/src/redis-server /usr/local/redis-3.0.6/redis.conf /etc/redis/7000
[root@andoncentos redis-3.0.6]# vim /etc/redis/7000/redis.conf
port 7000
daemonize yes
pidfile /var/run/redis_7000.pid
cluster-enabled yes
cluster-config-file nodes.conf
logfile "/var/log/redisd7000.log"
dir /etc/redis/7000/
cluster-node-timeout 5000
appendonly yes
4、修改redis服务的启动脚本,修改内容,并复制相关其他的节点配置
[root@andoncentos redis-3.0.6]# cp /usr/local/redis-3.0.6/utils/redis_init_script /etc/init.d/redis7000
[root@andoncentos redis-3.0.6]# vim /etc/init.d/redis7000
#!/bin/sh
# chkconfig 2345 90 10
# description:Redis is a persistent key-value database
# Simple Redis init.d script conceived to work on Linux systems
# as it does use of the /proc filesystem.
REDISPORT=7000
# EXEC=/usr/local/bin/redis-server
EXEC=/etc/redis/${REDISPORT}/redis-server
CLIEXEC=/usr/local/bin/redis-cli
PIDFILE=/var/run/redis_${REDISPORT}.pid
CONF="/etc/redis/${REDISPORT}/redis.conf"
[root@andoncentos redis-3.0.6]# cp /etc/init.d/redis7000 /etc/init.d/redis7001
[root@andoncentos redis-3.0.6]# cp /etc/init.d/redis7000 /etc/init.d/redis7003
[root@andoncentos redis-3.0.6]# cp /etc/init.d/redis7000 /etc/init.d/redis7004
5、设置为开机自启动服务器
[root@andoncentos redis-3.0.6]# chkconfig redis7000 on
[root@andoncentos redis-3.0.6]# chkconfig redis7001 on
[root@andoncentos redis-3.0.6]# chkconfig redis7003 on
[root@andoncentos redis-3.0.6]# chkconfig redis7004 on
6、重启系统,并检查redis7000,redis7001,redis7003,redis7004服务情况
[root@andoncentos redis-3.0.6]# reboot
[root@andoncentos 桌面]# systemctl status redis7004.service
(三)配置Redis集群
1、按照 ruby tree 工具,因为redis集群需要ruby
[root@andoncentos redis-3.0.6]# yum -y install tcl ruby tree
[root@andoncentos 桌面]# gem install redis --version 3.0.6
Fetching: redis-3.0.6.gem (100%)
Successfully installed redis-3.0.6
Parsing documentation for redis-3.0.6
Installing ri documentation for redis-3.0.6
1 gem installed
2、redis-trib.rb 配置集群
[root@andoncentos 桌面]# /usr/local/redis-3.0.6/src/redis-trib.rb create --replicas 1 192.168.1.110:7000 192.168.1.110:7001 192.168.1.110:7002 192.168.1.110:7003 192.168.1.110:7004 192.168.1.110:7005
>>> Creating cluster
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
192.168.1.110:7000
192.168.1.110:7001
192.168.1.110:7002
Adding replica 192.168.1.110:7003 to 192.168.1.110:7000
Adding replica 192.168.1.110:7004 to 192.168.1.110:7001
Adding replica 192.168.1.110:7005 to 192.168.1.110:7002
M: b164701893bfbdc078e2f7e3b16f1216c1bf65ff 192.168.1.110:7000
slots:0-5460 (5461 slots) master
M: 4c2d36c55cff692a7bbeccb663197b555747d15d 192.168.1.110:7001
slots:5461-10922 (5462 slots) master
M: b147e4dfcd63c5ce059540db55a9d7cb9fa093eb 192.168.1.110:7002
slots:10923-16383 (5461 slots) master
S: 757381aa5cc5c8ba70f3798f6de6cb7b2e97f924 192.168.1.110:7003
replicates b164701893bfbdc078e2f7e3b16f1216c1bf65ff
S: fecc8edf32fc72cd4a5d8ae5306fe4083abfe8e9 192.168.1.110:7004
replicates 4c2d36c55cff692a7bbeccb663197b555747d15d
S: 98bd8e1aff631a3bee7f92a39764decea16ee955 192.168.1.110:7005
replicates b147e4dfcd63c5ce059540db55a9d7cb9fa093eb
Can I set the above configuration? (type 'yes' to accept): <strong>yes</strong>
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join....
>>> Performing Cluster Check (using node 192.168.1.110:7000)
M: b164701893bfbdc078e2f7e3b16f1216c1bf65ff 192.168.1.110:7000
slots:0-5460 (5461 slots) master
M: 4c2d36c55cff692a7bbeccb663197b555747d15d 192.168.1.110:7001
slots:5461-10922 (5462 slots) master
M: b147e4dfcd63c5ce059540db55a9d7cb9fa093eb 192.168.1.110:7002
slots:10923-16383 (5461 slots) master
M: 757381aa5cc5c8ba70f3798f6de6cb7b2e97f924 192.168.1.110:7003
slots: (0 slots) master
replicates b164701893bfbdc078e2f7e3b16f1216c1bf65ff
M: fecc8edf32fc72cd4a5d8ae5306fe4083abfe8e9 192.168.1.110:7004
slots: (0 slots) master
replicates 4c2d36c55cff692a7bbeccb663197b555747d15d
M: 98bd8e1aff631a3bee7f92a39764decea16ee955 192.168.1.110:7005
slots: (0 slots) master
replicates b147e4dfcd63c5ce059540db55a9d7cb9fa093eb
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
3、检查集群状态
[root@andoncentos 桌面]# /usr/local/redis-3.0.6/src/redis-trib.rb check 192.168.1.110:7000
4、若出错的话,通过如下命令行修复
[root@andoncentos 桌面]# /usr/local/redis-3.0.6/src/redis-trib.rb fix 192.168.1.110:7000
5、防火墙开放端口,并重启防火墙
[root@andoncentos 桌面]# firewall-cmd --zone=public --add-port=7000-7005/tcp --permanent
success
[root@andoncentos 桌面]# firewall-cmd --reload
success
6、检查集群情况
[root@andoncentos 桌面]# redis-cli -c -p 7000
127.0.0.1:7000> cluster info
cluster_state:ok
cluster_slots_assigned:16384
cluster_slots_ok:16384
cluster_slots_pfail:0
cluster_slots_fail:0
cluster_known_nodes:6
cluster_size:3
cluster_current_epoch:6
cluster_my_epoch:1
cluster_stats_messages_sent:2492
cluster_stats_messages_received:2492
四、使用说明
(一)使用命令行测试缓存
(二)通过C#代码测试缓存
1、通过使用stackexchang.redis组件,将数据缓存到集群redis服务。
var cfg = RedisCachingSectionHandler.GetConfig();
var serializer = new NewtonsoftSerializer();
var redis = new StackExchangeRedisCacheClient(serializer, cfg);
var cls = new Cls(){ ID = 1, Name = txt };
string key = "tkey" + new Random().Next(1000, 9999).ToString();
redis.Add<Cls>(key, cls);
2、通过key获取数据
var cfg = RedisCachingSectionHandler.GetConfig();
var serializer = new NewtonsoftSerializer();
var redis = new StackExchangeRedisCacheClient(serializer, cfg);
var entity = redis.Get<Cls>(key);
ViewBag.KV = entity != null ? entity.Name : "";
3、redis缓存情况
五、总结
redis cluster 默认支持HA,但是对于单个Instance故障,使用者需要自行判断处理的机制,后续有待于研究,但是redis codis 这方面提供了方便的支持。