redis-cluster-proxy代理集群并支持mget命令

参考文档

搭建集群：http://www.redis.cn/topics/cluster-tutorial.html

redis-cluster-proxy：https://github.com/RedisLabs/redis-cluster-proxy

 

搭建集群

集群最少为3个节点，所以需要复制3份redis配置文件，每个文件更改下面几个地方：

• cluster-enabled：yes
• port：端口号
• pidfile：/var/run/redis-6379.pid
• logfile："/usr/local/var/log/redis/redis-6379.log"
• dir：/usr/local/var/db/redis/redis-6379

注意上面三个目录都需要给够写权限

$redis-server ./redis_6379.conf
$redis-server ./redis_6378.conf
$redis-server ./redis_6377.conf

$ps -ef | grep redis-server
  501 40954     1   0  3:53下午 ??         0:47.00 redis-server 127.0.0.1:6379 [cluster] 
  501 40956     1   0  3:53下午 ??         0:45.68 redis-server 127.0.0.1:6378 [cluster] 
  501 41001     1   0  3:57下午 ??         0:45.08 redis-server 127.0.0.1:6377 [cluster] 

创建集群

redis-cli --cluster create 127.0.0.1:6377 127.0.0.1:6378 127.0.0.1:6379

如果遇到错误：CLUSTERDOWN Hash slot not served，则需要修复错误中指定的实例的集群slot

$redis-cli --cluster fix 127.0.0.1:6379

如果遇到错误：[ERR] Node is not empty. Either the node already knows other nodes ，则需要清空对应节点的数据，并将持久化文件删除

redis-cli -p 6379 flushdb

成功

$redis-cli --cluster create 127.0.0.1:6377 127.0.0.1:6378 127.0.0.1:6379
>>> Performing hash slots allocation on 3 nodes...
Master[0] -> Slots 0 - 5460
Master[1] -> Slots 5461 - 10922
Master[2] -> Slots 10923 - 16383
M: 60e52412e798c15a8596490b24e81bc67e1e8b98 127.0.0.1:6377
   slots:[0-5460] (5461 slots) master
M: 0b9545dfaf4eb476755699f287e7187daa5eae05 127.0.0.1:6378
   slots:[5461-10922] (5462 slots) master
M: 781384b4ac76cad9f8a2b7a6c82afb0afdfa2a54 127.0.0.1:6379
   slots:[10923-16383] (5461 slots) master
Can I set the above configuration? (type 'yes' to accept): yes
>>> 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 127.0.0.1:6377)
M: 60e52412e798c15a8596490b24e81bc67e1e8b98 127.0.0.1:6377
   slots:[0-5460] (5461 slots) master
M: 781384b4ac76cad9f8a2b7a6c82afb0afdfa2a54 127.0.0.1:6379
   slots:[10923-16383] (5461 slots) master
M: 0b9545dfaf4eb476755699f287e7187daa5eae05 127.0.0.1:6378
   slots:[5461-10922] (5462 slots) master
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

 

通过 redis-cluster-proxy 启动代理

更改以下配置：

entry-point 127.0.0.1:6377
entry-point 127.0.0.1:6378
entry-point 127.0.0.1:6379

port 7777

bind 0.0.0.0

threads 8

daemonize yes

enable-cross-slot yes #开启跨node查询

$redis-cluster-proxy -c ./proxy.conf 

$ps -ef | grep redis
  501 43215     1   0  8:57下午 ??         0:02.90 redis-server 127.0.0.1:6377 [cluster] 
  501 43217     1   0  8:57下午 ??         0:02.89 redis-server 127.0.0.1:6379 [cluster] 
  501 43219     1   0  8:57下午 ??         0:02.89 redis-server 127.0.0.1:6378 [cluster] 
  501 43383     1   0  9:07下午 ??         0:00.03 redis-cluster-proxy -c ./proxy.conf

 

代码测试

此时可以支持普通集群不支持的mget等命令获取不同节点上的数据，proxy将汇总多个命令的结果。

先确认以下数据分布在不同节点上：

$redis-cli -p 6379
127.0.0.1:6379> keys *
1) "why"


$redis-cli -p 6378 
127.0.0.1:6378> keys *
(empty list or set)


$redis-cli -p 6377
127.0.0.1:6377> keys *
1) "jzm"

代码如下：

connect('127.0.0.1', 7777, 10);

$redis->set('why', 'why');
$redis->set('jzm', 'jzm');
$res = $redis->mget(['why', 'jzm']);

var_dump($res);die;

结果：

$php redis-cluster-proxy.php 
array(2) {
  [0]=>
  string(3) "why"
  [1]=>
  string(3) "jzm"
}

如果 enable-cross-slot 设置为默认的no，则返回false

$php redis-cluster-proxy.php 
bool(false)