周氏一族，整理技术文档，给下一代留点教程......
redis 3.0 cluster 安装篇，请看 http://zhoushouby.blog.51cto.com/9150272/1560400
本篇，是在 "redis 3.0 cluster 集群 学习之路篇 [2]" 的基础上面实现 “实际生产中Redis服务器部署” 安装调试以及测试。
六台服务器
server21 10.1.3.21
server22 10.1.3.22
server23 10.1.3.23
server24 10.1.3.24
server25 10.1.3.25
server26 10.1.3.26
---------------------------------每一台/etc/hosts里面都有添加一一对应记录
redis.conf 配置如下
port 6379 daemonize yes cluster-enabled yes cluster-config-file /usr/local/redis/nodes.conf cluster-node-timeout 5000 dir /usr/local/redis |
细心看，你会发现，上面多了一个路径，对的，个人习惯，我给redis创建了一个单独的目录，方便以后查找。
六台机器，所有配置一模一样，统一配置。
一路下来，完全没问题，每一台都能独立的运行redis服务
redis-server /usr/local/redis/redis.conf > /usr/local/redis/redis.log 2>&1 & |
但是，当我要来给他们开始集群的时候，就发生下面这么一幕情况
[root@server21 src]# ./redis-trib.rb create --replicas 1 10.1.3.21:6379 10.1.3.22:6379 10.1.3.26:6379 10.1.3.25:6379 10.1.3.24:6379 10.1.3.23:6379 >>> Creating cluster Connecting to node 10.1.3.21:6379: OK Connecting to node 10.1.3.22:6379: OK Connecting to node 10.1.3.26:6379: OK Connecting to node 10.1.3.25:6379: OK Connecting to node 10.1.3.24:6379: OK Connecting to node 10.1.3.23:6379: OK >>> Performing hash slots allocation on 6 nodes... Using 3 masters: 10.1.3.23:6379 10.1.3.24:6379 10.1.3.25:6379 Adding replica 10.1.3.26:6379 to 10.1.3.23:6379 Adding replica 10.1.3.22:7379 to 10.1.3.24:6379 Adding replica 10.1.3.21:7379 to 10.1.3.25:6379 S: fe32667413d75aa9dcebf28aa53fe4ed0df91bba 10.1.3.21:6379 replicates 7f4c3a0ea64c9f85e7d239c8968c94818decb9a8 S: 3b10b61ebf7ff4e618a945aa353e9e113541557a 10.1.3.22:6379 replicates f6c6fac2ada651819fac2ddc0daa91f54062a261 S: 0f9a72b4cb6b1ec2d20cc5f0d82be29925c8a050 10.1.3.26:6379 replicates 96dda21bf7dc9ba81c160375b39a59aedf78733d M: 7f4c3a0ea64c9f85e7d239c8968c94818decb9a8 10.1.3.25:6379 slots:10923-16383 (5461 slots) master M: f6c6fac2ada651819fac2ddc0daa91f54062a261 10.1.3.24:6379 slots:0-10922 (10923 slots) master M: 96dda21bf7dc9ba81c160375b39a59aedf78733d 10.1.3.23:6379 slots:0-10922 (10923 slots) master Can I set the above configuration? (type 'yes' to accept): yes /usr/lib64/ruby/gems/1.8/gems/redis-3.0.7/lib/redis/client.rb:97:in `call': ERR Slot 16011 is already busy (Redis::CommandError) from /usr/lib64/ruby/gems/1.8/gems/redis-3.0.7/lib/redis.rb:2432:in `method_missing' from /usr/lib64/ruby/gems/1.8/gems/redis-3.0.7/lib/redis.rb:37:in `synchronize' from /usr/lib/ruby/1.8/monitor.rb:242:in `mon_synchronize' from /usr/lib64/ruby/gems/1.8/gems/redis-3.0.7/lib/redis.rb:37:in `synchronize' from /usr/lib64/ruby/gems/1.8/gems/redis-3.0.7/lib/redis.rb:2431:in `method_missing' from ./redis-trib.rb:203:in `flush_node_config' from ./redis-trib.rb:629:in `flush_nodes_config' from ./redis-trib.rb:628:in `each' from ./redis-trib.rb:628:in `flush_nodes_config' from ./redis-trib.rb:964:in `create_cluster_cmd' from ./redis-trib.rb:1340:in `send' from ./redis-trib.rb:1340 [root@server21 src]# |
人生，从此不再寂寞了，开启蛋疼模式
开始找问题，蛋疼的就是他一个错误日志都没有，没法下手。
没办法之下，只好问度娘了，但是，网上看得我眼花缭乱，在大陆，只能找到这么一出玩意
http://blog.csdn.net/moxiaomomo/article/details/19333547
感觉有点道理，因为我的生产环境，之前是redis 2.x的版本，今天更新到3.0，可能会是我redis没卸载干净导致的。
好，二话不说，把服务器再清理一遍，然后重启一下。因为安装步骤都是按照我之前的来，没出现过问题，而且之前在VM里面装的时候，VM的系统是全新的，所以觉得他这个有点道理。开始捣腾起来了。
半个小时候过后，问题依旧，又来到这个问题了。
最终没办法，翻一下墙，泡一下老外redis开源问答区，碰见这个问题的，还不少，比中国大陆好多了，至少搜索下来，有那么10几篇文章，都是在咨询这个问题，但是没有最终的答案
别人不行，那就只能靠自己了。
/usr/lib64/ruby/gems/1.8/gems/redis-3.0.7/lib/redis/client.rb:97:
他提示我说97行这里不行，那就编辑他，gg到87行这里定位一下
终于，总算，找到问题所在了。这个是因为port和nodes文件的问题，不能设置成相同nodes名字的，而且端口也设置成不一样。
晒一下，我最终的配置
[root@server21 ~]# ll /usr/local/redis/ total 60 -rw-r--r-- 1 root root 89 Oct 3 20:02 appendonly.aof -rw-r--r-- 1 root root 18 Oct 3 20:17 dump.rdb -rw-r--r-- 1 root root 733 Oct 3 19:30 nodes-8021.conf -rw-r--r-- 1 root root 115 Oct 3 17:27 nodes.conf -rw-r--r-- 1 root root 37019 Oct 3 18:48 redis.conf -rw-r--r-- 1 root root 22 Oct 3 18:56 redis.log [root@server21 ~]# |
server21 port 8021 nodes-8021.conf
server22 port 8022 nodes-8022.conf
server23 port 8023 nodes-8023.conf
server24 port 8024 nodes-8024.conf
server25 port 8025 nodes-8025.conf
server26 port 8026 nodes-8026.conf
折腾了快两个小时，最后一次通过，这种感觉超爽!!!
[root@server21 src]# ./redis-trib.rb create --replicas 1 10.1.3.21:8021 10.1.3.22:8022 10.1.3.23:8023 10.1.3.24:8024 10.1.3.25:8025 10.1.3.26:8026 >>> Creating cluster Connecting to node 10.1.3.21:8021: OK Connecting to node 10.1.3.22:8022: OK Connecting to node 10.1.3.23:8023: OK Connecting to node 10.1.3.24:8024: OK Connecting to node 10.1.3.25:8025: OK Connecting to node 10.1.3.26:8026: OK >>> Performing hash slots allocation on 6 nodes... Using 3 masters: 10.1.3.26:8026 10.1.3.25:8025 10.1.3.24:8024 Adding replica 10.1.3.23:8023 to 10.1.3.26:8026 Adding replica 10.1.3.22:8022 to 10.1.3.25:8025 Adding replica 10.1.3.21:8021 to 10.1.3.24:8024 S: 690bb1a31fff14557483e1e6dea60e5e748fff41 10.1.3.21:8021 replicates 89efca75e220cdeb159e1e4954b953337563683f S: ae258d0a8e31d2cac64aaad62327ec85c3ce12d7 10.1.3.22:8022 replicates da00c6e9cf5f572396d7cfb006cbcb98e69928ee S: 8d45084d0ec34c4547c7e57228a5db0ba7eeeba2 10.1.3.23:8023 replicates e25ffd0663bf13d5b6aa6a397ffb0c21e4b11e05 M: 89efca75e220cdeb159e1e4954b953337563683f 10.1.3.24:8024 slots:10923-16383 (5461 slots) master M: da00c6e9cf5f572396d7cfb006cbcb98e69928ee 10.1.3.25:8025 slots:5461-10922 (5462 slots) master M: e25ffd0663bf13d5b6aa6a397ffb0c21e4b11e05 10.1.3.26:8026 slots:0-5460 (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 10.1.3.21:8021) M: 690bb1a31fff14557483e1e6dea60e5e748fff41 10.1.3.21:8021 slots: (0 slots) master replicates 89efca75e220cdeb159e1e4954b953337563683f M: ae258d0a8e31d2cac64aaad62327ec85c3ce12d7 10.1.3.22:8022 slots: (0 slots) master replicates da00c6e9cf5f572396d7cfb006cbcb98e69928ee M: 8d45084d0ec34c4547c7e57228a5db0ba7eeeba2 10.1.3.23:8023 slots: (0 slots) master replicates e25ffd0663bf13d5b6aa6a397ffb0c21e4b11e05 M: 89efca75e220cdeb159e1e4954b953337563683f 10.1.3.24:8024 slots:10923-16383 (5461 slots) master M: da00c6e9cf5f572396d7cfb006cbcb98e69928ee 10.1.3.25:8025 slots:5461-10922 (5462 slots) master M: e25ffd0663bf13d5b6aa6a397ffb0c21e4b11e05 10.1.3.26:8026 slots:0-5460 (5461 slots) master [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. [root@server21 src]# ./redis-trib.rb check 10.1.3.21:8021 Connecting to node 10.1.3.21:8021: OK Connecting to node 10.1.3.23:8023: OK Connecting to node 10.1.3.26:8026: OK Connecting to node 10.1.3.25:8025: OK Connecting to node 10.1.3.24:8024: OK Connecting to node 10.1.3.22:8022: OK >>> Performing Cluster Check (using node 10.1.3.21:8021) S: 690bb1a31fff14557483e1e6dea60e5e748fff41 10.1.3.21:8021 slots: (0 slots) slave replicates 89efca75e220cdeb159e1e4954b953337563683f S: 8d45084d0ec34c4547c7e57228a5db0ba7eeeba2 10.1.3.23:8023 slots: (0 slots) slave replicates e25ffd0663bf13d5b6aa6a397ffb0c21e4b11e05 M: e25ffd0663bf13d5b6aa6a397ffb0c21e4b11e05 10.1.3.26:8026 slots:0-5460 (5461 slots) master 1 additional replica(s) M: da00c6e9cf5f572396d7cfb006cbcb98e69928ee 10.1.3.25:8025 slots:5461-10922 (5462 slots) master 1 additional replica(s) M: 89efca75e220cdeb159e1e4954b953337563683f 10.1.3.24:8024 slots:10923-16383 (5461 slots) master 1 additional replica(s) S: ae258d0a8e31d2cac64aaad62327ec85c3ce12d7 10.1.3.22:8022 slots: (0 slots) slave replicates da00c6e9cf5f572396d7cfb006cbcb98e69928ee [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. [root@server21 src]# |
写下来，给后来者留点脚印......
