1
2
3
4
5
6
7
8
9

# Ubuntu/Debian
sudo apt install keepalived

# CentOS/RHEL
sudo dnf install keepalived

# 启动
sudo systemctl enable keepalived
sudo systemctl start keepalived

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

# /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {
   router_id LVS_DEVEL
   vrrp_skip_check_adv_addr
   vrrp_garp_interval 0
   vrrp_gna_interval 0
}

vrrp_instance VI_1 {
    state MASTER              # 初始状态：MASTER 或 BACKUP
    interface eth0            # 网卡名称
    virtual_router_id 51      # VRRP 路由 ID（同一组要相同）
    priority 100              # 优先级（MASTER 要比 BACKUP 高）
    advert_int 1              # 心跳间隔（秒）
    nopreempt                 # 非抢占模式
    
    authentication {
        auth_type PASS        # 认证类型
        auth_pass 1111        # 认证密码
    }
    
    virtual_ipaddress {
        192.168.1.100         # 虚拟 IP（VIP）
    }
    
    # 通知脚本
    notify_master "/etc/keepalived/notify.sh master"
    notify_backup "/etc/keepalived/notify.sh backup"
    notify_fault "/etc/keepalived/notify.sh fault"
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18

# /etc/keepalived/keepalived.conf

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 100
    advert_int 1
    
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    
    virtual_ipaddress {
        192.168.1.100
    }
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18

# /etc/keepalived/keepalived.conf

vrrp_instance VI_1 {
    state BACKUP
    interface eth0
    virtual_router_id 51
    priority 90               # 低于 MASTER
    advert_int 1
    
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    
    virtual_ipaddress {
        192.168.1.100
    }
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

# 检查脚本
vrrp_script chk_nginx {
    script "/usr/bin/pgrep nginx"
    interval 2                # 每2秒检查一次
    weight -20                # 检查失败时优先级减少20
    fall 2                    # 连续2次失败才算失败
    rise 1                    # 连续1次成功就算成功
}

vrrp_instance VI_1 {
    # ... 其他配置 ...
    
    track_script {
        chk_nginx            # 使用上面的检查脚本
    }
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

# /etc/keepalived/keepalived.conf

global_defs {
   router_id NGINX_MASTER
}

vrrp_script chk_nginx {
    script "pgrep nginx"
    interval 2
    weight -20
    fall 2
    rise 1
}

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 100
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    
    virtual_ipaddress {
        192.168.1.100/24 dev eth0
    }
    
    track_script {
        chk_nginx
    }
}

1

# /etc/keepalived/keepalived.conf（除了 priority 90，其他相同）

1
2
3
4
5

# 简单认证
authentication {
    auth_type PASS
    auth_pass 1111
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

# CentOS/RHEL (需要 EPEL)
sudo yum install pacemaker pcs

# Ubuntu/Debian
sudo apt install pacemaker pcs

# 启动并设置开机启动
sudo systemctl start pcsd
sudo systemctl enable pcsd

# 设置 hacluster 密码
sudo passwd hacluster

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

# 1. 节点认证
sudo pcs host auth node1 node2

# 2. 创建集群（老写法是 pcs cluster setup my_cluster --start node1 node2，
#    pcs 0.10 起 --start 已被移除，改为 setup 之后再单独 start）
sudo pcs cluster setup my_cluster node1 node2

# 3. 启用集群
sudo pcs cluster start --all
sudo pcs cluster enable --all

# 4. 检查状态
sudo pcs cluster status

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

# 创建资源（VIP）
sudo pcs resource create VIP ocf:heartbeat:IPaddr2 \
    ip=192.168.1.100 \
    cidr_netmask=24 \
    op monitor interval=30s

# 创建资源（Nginx）
sudo pcs resource create WebServer systemd:nginx

# 创建资源（MySQL）
sudo pcs resource create MySQL ocf:heartbeat:mysql \
    binary=/usr/bin/mysqld \
    config=/etc/my.cnf \
    datadir=/var/lib/mysql \
    op monitor interval=20s

# 查看资源
sudo pcs resource config        # pcs 0.10 起推荐用它；pcs resource show 已标记为过时

# 启动资源
sudo pcs resource start VIP

# 资源约束
sudo pcs constraint colocation add WebServer VIP INFINITY
sudo pcs constraint order start VIP then WebServer

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

# 手动迁移资源
sudo pcs resource move WebServer node2

# 查看资源位置（注意：上面这条 move 会悄悄生成一条 -INFINITY 的位置约束，
# 不清理的话资源就再也回不来了，故障恢复后一定要执行 clear）
sudo pcs resource clear WebServer

# 查看资源位置
sudo pcs resource status WebServer

# 设置资源粘性（偏好当前节点）
sudo pcs resource meta WebServer resource-stickiness=100

# 亲和性约束
sudo pcs constraint location WebServer prefers node1=200
sudo pcs constraint location WebServer prefers node2=100

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

# 查看集群状态
sudo pcs status

# 查看资源详情
sudo pcs resource show VIP --full

# 查看事件日志
sudo journalctl -u pacemaker -f

# 查看资源代理
sudo pcs resource standards
sudo pcs resource agents ocf:heartbeat

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

# /etc/keepalived/keepalived.conf (MASTER)

global_defs {
   router_id lb-master
}

vrrp_script chk_nginx {
    script "/etc/keepalived/check_nginx.sh"
    interval 2
    weight -20
}

vrrp_instance VI_1 {
    state MASTER
    interface eth0
    virtual_router_id 51
    priority 100
    advert_int 1
    
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    
    virtual_ipaddress {
        192.168.1.100/24
    }
    
    track_script {
        chk_nginx
    }
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

# /etc/keepalived/check_nginx.sh
#!/bin/bash

# 退出码 0 表示"健康"，非 0 表示"异常"，keepalived 只看退出码

# 第一层：进程还在不在（比 pgrep 更规范，能识别 systemd 管理的服务状态）
systemctl is-active --quiet nginx || exit 1

# 第二层：服务是否真的能响应请求（进程活着但卡死时，上面那条检查不出来）
curl -fsS -m 2 -o /dev/null http://127.0.0.1/healthz || exit 1

exit 0

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20

# Ceph 架构
ceph osd tree

# 集群整体健康状态（先看这一条，再往下查）
ceph -s

# 查看 MON（监控）/ MGR（管理）/ OSD（存储）组件分布
ceph mon stat
ceph mgr dump | head -20

# 副本池配置
ceph osd pool set mypool size 3
ceph osd pool set mypool min_size 2

# 查看池的副本设置
ceph osd pool get mypool size
ceph osd pool get mypool min_size

# PG 状态
ceph pg stat

1
2
3
4
5
6

# 主备切换场景
# 1. 主服务器故障检测
# 2. Keepalived 释放 VIP
# 3. 备服务器接管 VIP
# 4. 启动服务
# 5. 用户无感知恢复

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

# 1. 配置主从复制
# 主库 my.cnf
[mysqld]
server-id=1
log-bin=mysql-bin
binlog-format=ROW

# 从库 my.cnf
[mysqld]
server-id=2
relay-log=relay-bin
read-only=1

# 2. 主库创建复制用户
# 注意：MySQL 8 已不支持在 GRANT 里顺带创建用户并设密码（老写法会直接报语法错误），
# 必须分两步：先 CREATE USER，再 GRANT 权限
CREATE USER 'repl'@'%' IDENTIFIED BY 'StrongPassword!';
GRANT REPLICATION SLAVE ON *.* TO 'repl'@'%';

# 3. 从库配置主库信息
# MySQL 8.0.23 起 CHANGE MASTER TO 已废弃，改用 CHANGE REPLICATION SOURCE TO
CHANGE REPLICATION SOURCE TO
    SOURCE_HOST='主库IP',
    SOURCE_USER='repl',
    SOURCE_PASSWORD='StrongPassword!',
    SOURCE_LOG_FILE='mysql-bin.000001',
    SOURCE_LOG_POS=123;

# 4. 启动复制并查看状态
START REPLICA;
SHOW REPLICA STATUS\G

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

# 安装 MHA
dnf install mha4mysql-node mha4mysql-manager

# 配置 MHA
# /etc/app1.cnf
[server default]
user=mha
password=mha
manager_workdir=/var/log/mha
remote_workdir=/var/log/mha

[server1]
hostname=192.168.1.101

[server2]
hostname=192.168.1.102
candidate_master=1

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

# 安装 Galera
# CentOS
yum install MariaDB-server-galera

# 配置 Galera
# /etc/my.cnf.d/server.cnf
[galera]
wsrep_on=ON
wsrep_cluster_name="my_cluster"
wsrep_cluster_address="gcomm://192.168.1.101,192.168.1.102,192.168.1.103"
wsrep_node_address=192.168.1.101
wsrep_provider=/usr/lib/galera/libgalera_smm.so

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15

# 1. 启动主从
redis-server --port 6379 --daemonize yes
redis-server --port 6380 --daemonize yes --slaveof 127.0.0.1 6379
# 说明：slaveof 是旧词，新版本更推荐 replicaof，两者目前都还能用
# redis-server --port 6380 --daemonize yes --replicaof 127.0.0.1 6379

# 2. 启动 Sentinel
redis-sentinel /etc/sentinel.conf

# 3. Sentinel 配置
# /etc/sentinel.conf
sentinel monitor mymaster 127.0.0.1 6379 2
sentinel down-after-milliseconds mymaster 5000
sentinel parallel-syncs mymaster 1
sentinel failover-timeout mymaster 900000

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

# 创建集群
# 注意：--cluster-replicas 1 表示"每个主节点配 1 个副本"，
# 因此 3 主 3 从至少需要 6 个节点；只写 3 个节点会直接报错
# "Invalid configuration for cluster creation. Redis Cluster requires at least 3 master nodes."
redis-cli --cluster create \
    192.168.1.101:6379 192.168.1.102:6379 192.168.1.103:6379 \
    192.168.1.104:6379 192.168.1.105:6379 192.168.1.106:6379 \
    --cluster-replicas 1

# 如果只是实验环境、没有 6 台机器，可以明确起见把副本数设为 0
redis-cli --cluster create \
    192.168.1.101:6379 192.168.1.102:6379 192.168.1.103:6379 \
    --cluster-replicas 0

# 查看集群状态
redis-cli -c cluster info
redis-cli -c cluster nodes

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19

# 1. 查看 Keepalived 状态
systemctl status keepalived
journalctl -u keepalived -f        # 实时看日志；Debian 系没有 /var/log/messages，用这个最稳

# 2. 查看 VRRP 状态
journalctl -u keepalived | grep -i vrrp        # RHEL 系也可以看 /var/log/messages

# 3. 检查 VIP 是否绑定
ip addr show | grep 192.168.1.100

# 4. 测试 VRRP 通信
tcpdump -n -i eth0 vrrp          # -n 避免反向解析，输出更干净
# 收不到任何 VRRP 报文 = 心跳不通，重点查：组播是否被禁（云环境改用 unicast_peer）、
# 防火墙是否放行了协议号 112、两台机器是否真在同一二层网络

# 5. 常见问题
# - VIP 没有绑定：检查防火墙、优先级
# -频繁切换：检查网络稳定性、心跳间隔
# - 服务未启动：检查 notify 脚本、优先级计算

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20

# 查看集群状态
sudo pcs cluster status

# 查看资源详细状态
sudo pcs resource show

# 查看约束
sudo pcs constraint show

# 设置资源亲和性（在一起）
sudo pcs constraint colocation add webserver with VIP INFINITY

# 设置启动顺序
sudo pcs constraint order VIP then webserver

# 设置资源粘性（倾向留在当前节点）
sudo pcs resource meta webserver resource-stickiness=100

# 测试故障转移
sudo pcs resource move webserver node2