（进阶篇）Redis6.2.0 集群哨兵模式_故障转移_03

￡神魔★判官ぃ 2022-10-12 12:21 150阅读 0赞

### 文章目录 ###

*   *   *   *   *  一、前置准备
                 *   *  启动redis服务
                     *  启动sentinel服务
                     *  查看sentiel启动日志
                 *  二、实战故障转移
                 *   *  2.1. 主节点停止
                     *  2.2. Sentienl监控日志
                     *  2.3. 节点信息监控
                     *  2.4. 故障转移后变化
                     *  2.5. 重新启动
                     *  2.6. sentine日志
                     *  2.7. 节点信息监控
                     *  2.8. 配置文件变化
                     *  2.9. 读写验证

##### 一、前置准备 #####

###### 启动redis服务 ######

# 查看redis进程
    ps -ef |grep redis
    
    # 启动redis服务
    /usr/local/redis/bin/redis-server /usr/local/redis/conf/redis.conf
    
    #查看主从节点信息
    /usr/local/redis/bin/redis-cli -a 123456
    info replication

###### 启动sentinel服务 ######

> 声明：在启动3个redis服务的前提下，再依次启动3个sentinel服务

/usr/local/redis/bin/redis-sentinel /usr/local/redis/conf/sentiel.conf

###### 查看sentiel启动日志 ######

tail -f /usr/local/redis/log/sentinel.log

##### 二、实战故障转移 #####

###### 2.1. 主节点停止 ######

[root@bigdata01 redis-6.2.0]# /usr/local/redis/bin/redis-cli -a 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> info replcation
    127.0.0.1:6379> info replication
    # Replication
    role:master
    connected_slaves:2
    slave0:ip=xxx.xxx.92.101,port=6379,state=online,offset=3xxx,lag=1
    slave1:ip=xxx.xxx.92.102,port=6379,state=online,offset=3xxx,lag=1
    master_failover_state:no-failover
    master_replid:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_replid2:0000000000000000000000000000000000000000
    master_repl_offset:3xxx
    second_repl_offset:-1
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1391
    repl_backlog_histlen:1778
    127.0.0.1:6379> SHUTDOWN
    not connected>

###### 2.2. Sentienl监控日志 ######

mymaster 主节点 xxx.xxx.92.100 6379 被添加主观下线标记，其他的sentinel每秒一次和xxx.xxx.92.100 6379的通信
    # +sdown master mymaster xxx.xxx.92.100 6379
    启动一个新选举流程，记录当前选举的次数
    # +new-epoch 1
    投票选举一个sentinel去做故障转移这件事
    # +vote-for-leader a20913bfcbeb3a5d8d58ce1bd372cc40455xxx10 1
    客观下线：当前环境内多个sentinel节点对我们的某一个服务器标记为主观下线，满足quorum 仲裁的选项值之后，标记为客观下线
    # +odown master mymaster xxx.xxx.92.100 6379 #quorum 3/2
    # Next failover delay: I will not start a failover before Sun Jul 4 22:39:56 2021
    # +config-update-from sentinel a20913bfcbeb3a5d8d58ce1bd372cc40455xxx10 xxx.xxx.92.1026379
    # +switch-master mymaster xxx.xxx.92.100 6379 xxx.xxx.92.101 6379
    * +slave slave xxx.xxx.92.102:6379 xxx.xxx.92.102 6379 @ mymaster xxx.xxx.92.101 6379
    * +slave slave xxx.xxx.92.100:6379 xxx.xxx.92.100 6379 @ mymaster xxx.xxx.92.101 6379
    # +sdown slave xxx.xxx.92.100:6379 xxx.xxx.92.100 6379 @ mymaster xxx.xxx.92.101 6379

###### 2.3. 节点信息监控 ######

查看故障转移后的节点信息  
01节点信息

02节点信息

###### 2.4. 故障转移后变化 ######

01节点由slave节点变成了master节点，02仍然后slave节点

###### 2.5. 重新启动 ######

故障的 100节点(原主节点)重新启动

/usr/local/redis/bin/redis-server /usr/local/redis/conf/redis.conf

###### 2.6. sentine日志 ######

* +convert-to-slave slave xxx.xxx.92.100:6379 xxx.xxx.92.100 6379 @ mymaster xxx.xxx.92.101 6379

###### 2.7. 节点信息监控 ######

[root@bigdata01 redis-6.2.0]# /usr/local/redis/bin/redis-cli -a 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> info replication
    # Replication
    role:slave
    master_host:xxx.xxx.92.101
    master_port:6379
    master_link_status:up
    master_last_io_seconds_ago:0
    master_sync_in_progress:0
    slave_repl_offset:528510
    slave_priority:100
    slave_read_only:1
    connected_slaves:0
    master_failover_state:no-failover
    master_replid:424069affa171e62af3e2c65a2ee1cb27ecb6c2f
    master_replid2:0000000000000000000000000000000000000000
    master_repl_offset:528510
    second_repl_offset:-1
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:518136
    repl_backlog_histlen:10375
    127.0.0.1:6379>

01节点信息

[root@bigdata02 ~]# /usr/local/redis/bin/redis-cli -a 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> info replication
    # Replication
    role:slave
    master_host:xxx.xxx.92.100
    master_port:6379
    master_link_status:up
    master_last_io_seconds_ago:6
    master_sync_in_progress:0
    slave_repl_offset:3196
    slave_priority:100
    slave_read_only:1
    connected_slaves:0
    master_failover_state:no-failover
    master_replid:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_replid2:0000000000000000000000000000000000000000
    master_repl_offset:3196
    second_repl_offset:-1
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1391
    repl_backlog_histlen:1806
    
    # 故障转移后 -------------------------------------------------------------------
    127.0.0.1:6379> info replication
    # Replication
    role:master
    connected_slaves:1
    slave0:ip=xxx.xxx.92.102,port=6379,state=online,offset=380674,lag=1
    master_failover_state:no-failover
    master_replid:424069affa171e62af3e2c65a2ee1cb27ecb6c2f
    master_replid2:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_repl_offset:380674
    second_repl_offset:156702
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1391
    repl_backlog_histlen:379284
    
    
    #原主节点重新启动后 -------------------------------------------------------------------
    127.0.0.1:6379> info replication
    # Replication
    role:master
    connected_slaves:2
    slave0:ip=xxx.xxx.92.102,port=6379,state=online,offset=537589,lag=1
    slave1:ip=xxx.xxx.92.100,port=6379,state=online,offset=537732,lag=0
    master_failover_state:no-failover
    master_replid:424069affa171e62af3e2c65a2ee1cb27ecb6c2f
    master_replid2:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_repl_offset:537875
    second_repl_offset:156702
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1391
    repl_backlog_histlen:536485
    127.0.0.1:6379>

02节点信息

[root@bigdata03 ~]# /usr/local/redis/bin/redis-cli -a 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> info replication
    # Replication
    role:slave
    master_host:xxx.xxx.92.100
    master_port:6379
    master_link_status:up
    master_last_io_seconds_ago:4
    master_sync_in_progress:0
    slave_repl_offset:3210
    slave_priority:100
    slave_read_only:1
    connected_slaves:0
    master_failover_state:no-failover
    master_replid:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_replid2:0000000000000000000000000000000000000000
    master_repl_offset:3210
    second_repl_offset:-1
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1433
    repl_backlog_histlen:1778
    
    # 故障转移后 -------------------------------------------------------------------
    127.0.0.1:6379> info replication
    # Replication
    role:slave
    master_host:xxx.xxx.92.101
    master_port:6379
    master_link_status:up
    master_last_io_seconds_ago:0
    master_sync_in_progress:0
    slave_repl_offset:381546
    slave_priority:100
    slave_read_only:1
    connected_slaves:0
    master_failover_state:no-failover
    master_replid:424069affa171e62af3e2c65a2ee1cb27ecb6c2f
    master_replid2:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_repl_offset:381546
    second_repl_offset:156702
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1433
    repl_backlog_histlen:380114
    
    #原主节点重新启动后 -------------------------------------------------------------------
    127.0.0.1:6379> info replication
    # Replication
    role:slave
    master_host:xxx.xxx.92.101
    master_port:6379
    master_link_status:up
    master_last_io_seconds_ago:0
    master_sync_in_progress:0
    slave_repl_offset:556005
    slave_priority:100
    slave_read_only:1
    connected_slaves:0
    master_failover_state:no-failover
    master_replid:424069affa171e62af3e2c65a2ee1cb27ecb6c2f
    master_replid2:4f07bacde969998356194ad698c6dd7e2f6f7da2
    master_repl_offset:556005
    second_repl_offset:156702
    repl_backlog_active:1
    repl_backlog_size:1048576
    repl_backlog_first_byte_offset:1433
    repl_backlog_histlen:554573
    127.0.0.1:6379>

###### 2.8. 配置文件变化 ######

原100主节点

# 放行访问IP的限制 多个用空格隔开
    # 0.0.0.0 代表无限制
    bind 0.0.0.0
    
    # 后台启动开启
    daemonize yes
    
    # 日志存储目录以及日志文件名
    logfile "/usr/local/redis/log/redis.log"
    
    # rdb数据文件名
    dbfilename "dump.rdb"
    
    # aof模式开启
    appendonly no
    
    # aof数据文件名
    appendfilename "appendonly.aof"
    
    #rdb数据文件和aof数据文件存储目录
    dir "/usr/local/redis/data"
    
    # 设置密码
    requirepass 123456
    
    # 从节点访问主节点的密码(必须和requirepass 一致)
    masterauth "123456"
    
    #从节点只读模式
    replica-read-only yes
    # Generated by CONFIG REWRITE
    pidfile "/var/run/redis.pid"
    save 3600 1
    save 300 100
    save 60 10000
    user default on #8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92 ~* &* +@all
    replicaof xxx.xxx.92.101 6379

变化：最下面添加了主节点信息

原01节点

# 放行访问IP的限制 多个用空格隔开
    # 0.0.0.0 代表无限制
    bind 0.0.0.0
    
    # 后台启动开启
    daemonize yes
    
    # 日志存储目录以及日志文件名
    logfile "/usr/local/redis/log/redis.log"
    
    # rdb数据文件名
    dbfilename "dump.rdb"
    
    # aof模式开启
    appendonly no
    
    # aof数据文件名
    appendfilename "appendonly.aof"
    
    #rdb数据文件和aof数据文件存储目录
    dir "/usr/local/redis/data"
    
    # 设置密码
    requirepass 123456
    
    # 从节点访问主节点的密码(必须和requirepass 一致)
    masterauth "123456"
    
    #从节点只读模式
    replica-read-only yes
    
    # 从节点从属于那个主节点
    
    # Generated by CONFIG REWRITE
    pidfile "/var/run/redis.pid"
    save 3600 1
    save 300 100
    save 60 10000
    user default on #8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92 ~* &* +@all

变化：原来配置的主节点信息没了

02节点信息

# 放行访问IP的限制 多个用空格隔开
    # 0.0.0.0 代表无限制
    bind 0.0.0.0
    
    # 后台启动开启
    daemonize yes
    
    # 日志存储目录以及日志文件名
    logfile "/usr/local/redis/log/redis.log"
    
    # rdb数据文件名
    dbfilename "dump.rdb"
    
    # aof模式开启
    appendonly no
    
    # aof数据文件名
    appendfilename "appendonly.aof"
    
    #rdb数据文件和aof数据文件存储目录
    dir "/usr/local/redis/data"
    
    # 设置密码
    requirepass 123456
    
    # 从节点访问主节点的密码(必须和requirepass 一致)
    masterauth "123456"
    
    #从节点只读模式
    replica-read-only yes
    
    # 从节点从属于那个主节点
    replicaof xxx.xxx.92.101 6379
    # Generated by CONFIG REWRITE
    pidfile "/var/run/redis.pid"
    save 3600 1
    save 300 100
    save 60 10000
    user default on #8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92 ~* &* +@all

###### 2.9. 读写验证 ######

# 登陆redis客户端
    /usr/local/redis/bin/redis-cli --user default --pass 123456

#故障转移后的100机器现在为slave节点，可读不可写
    [root@bigdata01 redis-6.2.0]# /usr/local/redis/bin/redis-cli --user default --pass 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> set name lisi
    (error) READONLY You can't write against a read only replica.
    127.0.0.1:6379>

#故障转移后的101机器现在为master节点，可以读写
    [root@bigdata02 ~]# /usr/local/redis/bin/redis-cli --user default --pass 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> set name lisi
    OK
    127.0.0.1:6379> get name
    "lisi"
    127.0.0.1:6379>

#故障转移后的102机器现在为slave节点，可读不可写
    [root@bigdata03 ~]# /usr/local/redis/bin/redis-cli --user default --pass 123456
    Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
    127.0.0.1:6379> set name wangwu 
    (error) READONLY You can't write against a read only replica.
    127.0.0.1:6379> get name
    "lisi"
    127.0.0.1:6379>