安装和配置MySQL Server和MySQL InnoDB群集
#devops #mysql #ansible

This角色将使用MySQL InnoDB Cluster或GTID replication

在HA模式下安装和配置MySQL Server或MySQL

目录

角色变量

此角色接受此变量：

var 检索收听IP地址
必需 默认 desc
mysql_subnet yes 192.168.25.0/24 子网在哪里聆听。如果VM或Bare Metal Server具有多个接口，则Ansible将过滤界面，而MySQL将仅在特定接口上收听。该变量还用于计算MySQL Server ID。
mysql_root_pw yes mysql root密码。
mysql_authentication no mysql_native_password mySQL身份验证方法。
disable_firewall no no 如果设置为Yes Ansible将禁用防火墙。
disable_selinux no no 禁用SELINUX。默认号，如果要配置Selinux，请使用另一个角色。您可以禁用SELINUX将此变量设置为Yes
resolv_mode no dns mySQL如何解析名称，默认DNS。如果设置为主机/etc/hosts文件将被覆盖
mysql_listen_all_interfaces no no 将此变量设置为YES，以允许MySQL在所有接口上收听0.0.0.0/0。否则，将使用 mysql_subnet variable
mysql_user no mysql mySQL系统用户
mysql_group no mysql MySQL搜索系统的组用户
mysql_data_dir no /var/lib/mysql mysql数据dir
mysql_log_dir no /var/log/mysql mysql log dir
mysql_conf_dir no /etc/mysql mysql conf dir
mysql_pid_dir no /var/run/mysqld mysql pid dir
mysql_operator_user no operator mySQL操作员用户，用于引导mysql innodb群集。
mysql_operator_password no Op3r4torMyPw 操作员用户的密码
mysql_replica_user no replica mySQL副本用户。用于所有复制操作
mysql_replica_password no rEpL1c4p4Sw0,rd 副本用户的密码
mysql_replication_mode no InnoDB Cluster，GTID，空/无（默认）
mysql_gr_name no 如果 mysql_replication_mode 设置为 innodb cluster 。组复制的UUID
mysql_gr_vcu no 如果 mysql_replication_mode 设置为 innodb cluster 。小组复制view change uuid
mysql_innodb_cluster_name no 如果 mysql_replication_mode 设置为 innodb cluster 。 mysql innodb群集的名称

流动，建立测试基础架构

要测试这个角色，我们使用Vagrant和Virtualbox，但是如果您愿意，您也可以使用自己的VMS或Baremetal机器。

第一步是下载this repo和birng所有VM。但首先在 change_me 变量中的vagrantfile粘贴您的public ssh键中。您还可以通过更改NNODES变量来调整部署的VM数量（在此Exaple中，我们将使用5个节点）。现在我们准备好配备机器：

git clone https://github.com/garutilorenzo/ansible-role-linux-mysql.git

cd ansible-role-linux-mysql/

vagrant up
Bringing machine 'my-ubuntu-0' up with 'virtualbox' provider...
Bringing machine 'my-ubuntu-1' up with 'virtualbox' provider...
Bringing machine 'my-ubuntu-2' up with 'virtualbox' provider...
Bringing machine 'my-ubuntu-3' up with 'virtualbox' provider...
Bringing machine 'my-ubuntu-4' up with 'virtualbox' provider...

[...]
[...]

    my-ubuntu-4: Inserting generated public key within guest...
==> my-ubuntu-4: Machine booted and ready!
==> my-ubuntu-4: Checking for guest additions in VM...
    my-ubuntu-4: The guest additions on this VM do not match the installed version of
    my-ubuntu-4: VirtualBox! In most cases this is fine, but in rare cases it can
    my-ubuntu-4: prevent things such as shared folders from working properly. If you see
    my-ubuntu-4: shared folder errors, please make sure the guest additions within the
    my-ubuntu-4: virtual machine match the version of VirtualBox you have installed on
    my-ubuntu-4: your host and reload your VM.
    my-ubuntu-4:
    my-ubuntu-4: Guest Additions Version: 6.0.0 r127566
    my-ubuntu-4: VirtualBox Version: 6.1
==> my-ubuntu-4: Setting hostname...
==> my-ubuntu-4: Configuring and enabling network interfaces...
==> my-ubuntu-4: Mounting shared folders...
    my-ubuntu-4: /vagrant => C:/Users/Lorenzo Garuti/workspaces/simple-ubuntu
==> my-ubuntu-4: Running provisioner: shell...
    my-ubuntu-4: Running: inline script
==> my-ubuntu-4: Running provisioner: shell...
    my-ubuntu-4: Running: inline script
    my-ubuntu-4: hello from node 5

Ansible设置和飞行前检查

现在，如果您没有安装易菜，请安装Ansible和所有要求：

apt-get install python3 python3-pip uuidgen openssl
pip3 install pipenv

pipenv shell
pip install -r requirements.txt

现在安装了Ansible，我们可以直接从GitHub下载该角色：

ansible-galaxy install git+https://github.com/garutilorenzo/ansible-role-linux-mysql.git

Whit Ansible和我们可以设置库存文件（hosts.ini）的角色：

[mysql]
my-ubuntu-0 ansible_host=192.168.25.110
my-ubuntu-1 ansible_host=192.168.25.111
my-ubuntu-2 ansible_host=192.168.25.112
my-ubuntu-3 ansible_host=192.168.25.113
my-ubuntu-4 ansible_host=192.168.25.114

和vars.yml文件：

---

disable_firewall: yes
disable_selinux: yes
mysql_resolv_mode: hosts
mysql_subnet: 192.168.25.0/24
mysql_listen_all_interfaces: yes

mysql_root_pw: '<CHANGE_ME>' # <- openssl rand -base64 32 | sed 's/=//'
mysql_replication_mode: 'InnoDB Cluster'
mysql_gr_name: '<CHANGE_ME>' # <- uuidgen
mysql_gr_vcu: '<CHANGE_ME>' #  <- uuidgen
mysql_innodb_cluster_name: 'cluster_lab'

注意 mysql_gr_name和mysql_gr_vcu是不同的uuid，所以两次运行uuidgen。
使用此Vars，我们将使用MySQL InnoDB群集在HA模式下部署MySQL，将从现有的Group Replication配置创建群集。

进行安装之前的最后一步是创建网站。 

---
- hosts: mysql
  become: yes
  remote_user: vagrant
  roles: 
    - role: ansible-role-linux-mysql
  vars_files:
    - vars.yml

部署mysql innodb群集

我们终于准备好使用Ansible部署MySQL InnoDB群集：

export ANSIBLE_HOST_KEY_CHECKING=False # Ansible skip ssh-key validation

ansible-playbook -i hosts.ini site.yml -e mysql_bootstrap_host=my-ubuntu-0

TASK [ansible-role-linux-mysql : render mysql.conf.d/mysqld.cnf] *******************************************************************************************
ok: [my-ubuntu-0]
ok: [my-ubuntu-2]
ok: [my-ubuntu-1]
ok: [my-ubuntu-3]
ok: [my-ubuntu-4]

TASK [ansible-role-linux-mysql : render mysql.conf.d/gtid.cnf] *********************************************************************************************
ok: [my-ubuntu-1]
ok: [my-ubuntu-0]
ok: [my-ubuntu-3]
ok: [my-ubuntu-2]
ok: [my-ubuntu-4]

TASK [ansible-role-linux-mysql : ansible.builtin.fail] *****************************************************************************************************
skipping: [my-ubuntu-0]
skipping: [my-ubuntu-1]
skipping: [my-ubuntu-2]
skipping: [my-ubuntu-3]
skipping: [my-ubuntu-4]

TASK [ansible-role-linux-mysql : ansible.builtin.fail] *****************************************************************************************************
skipping: [my-ubuntu-0]
skipping: [my-ubuntu-1]
skipping: [my-ubuntu-2]
skipping: [my-ubuntu-3]
skipping: [my-ubuntu-4]

TASK [ansible-role-linux-mysql : render innodb_cluster.cnf] ************************************************************************************************
ok: [my-ubuntu-0]
ok: [my-ubuntu-1]
ok: [my-ubuntu-3]
ok: [my-ubuntu-2]
ok: [my-ubuntu-4]

RUNNING HANDLER [ansible-role-linux-mysql : reload systemd] ************************************************************************************************
ok: [my-ubuntu-3]
ok: [my-ubuntu-0]
ok: [my-ubuntu-2]
ok: [my-ubuntu-4]
ok: [my-ubuntu-1]

PLAY RECAP *************************************************************************************************************************************************
my-ubuntu-0               : ok=69   changed=27   unreachable=0    failed=0    skipped=12   rescued=0    ignored=0   
my-ubuntu-1               : ok=71   changed=28   unreachable=0    failed=0    skipped=10   rescued=0    ignored=0   
my-ubuntu-2               : ok=71   changed=28   unreachable=0    failed=0    skipped=10   rescued=0    ignored=0   
my-ubuntu-3               : ok=71   changed=28   unreachable=0    failed=0    skipped=10   rescued=0    ignored=0   
my-ubuntu-4               : ok=71   changed=28   unreachable=0    failed=0    skipped=10   rescued=0    ignored=0

现在安装了群集，但是我们必须持续一些配置。由于群集是一个新的群集，因此Ansible已在bootstrap模式下启动了组复制。这意味着第一个实例（在这种情况下为my-ubuntu-0）具有te value group_replication_bootstrap_group_group 设置为 on 和 group_replication_group_seeds 空值。第二次运行的Ansible，此变量将设置为正确的值：

ansible-playbook -i hosts.ini site.yml

TASK [ansible-role-linux-mysql : ansible.builtin.fail] *****************************************************************************************************
skipping: [my-ubuntu-0]
skipping: [my-ubuntu-1]
skipping: [my-ubuntu-2]
skipping: [my-ubuntu-3]
skipping: [my-ubuntu-4]

TASK [ansible-role-linux-mysql : ansible.builtin.fail] *****************************************************************************************************
skipping: [my-ubuntu-0]
skipping: [my-ubuntu-1]
skipping: [my-ubuntu-2]
skipping: [my-ubuntu-3]
skipping: [my-ubuntu-4]

TASK [ansible-role-linux-mysql : render innodb_cluster.cnf] ************************************************************************************************
ok: [my-ubuntu-2]
ok: [my-ubuntu-4]
ok: [my-ubuntu-1]
ok: [my-ubuntu-3]
changed: [my-ubuntu-0]

PLAY RECAP *************************************************************************************************************************************************
my-ubuntu-0               : ok=30   changed=1    unreachable=0    failed=0    skipped=18   rescued=0    ignored=0   
my-ubuntu-1               : ok=30   changed=0    unreachable=0    failed=0    skipped=18   rescued=0    ignored=0   
my-ubuntu-2               : ok=30   changed=0    unreachable=0    failed=0    skipped=18   rescued=0    ignored=0   
my-ubuntu-3               : ok=30   changed=0    unreachable=0    failed=0    skipped=18   rescued=0    ignored=0   
my-ubuntu-4               : ok=30   changed=0    unreachable=0    failed=0    skipped=18   rescued=0    ignored=0

在本指南中，MySQLSH用于在MySQL InnoDB群集上进行操作。 Here您可以找到有关MySQLSH的更多信息。

现在我们最终可以检查我们的群集：

root@my-ubuntu-0:~# mysqlsh root@my-ubuntu-0
Please provide the password for 'root@my-ubuntu-0': ******************************************
MySQL  localhost:33060+ ssl  JS > clu = dba.getCluster()
MySQL  localhost:33060+ ssl  JS > clu.status()
{
    "clusterName": "cluster_lab", 
    "defaultReplicaSet": {
        "name": "default", 
        "primary": "my-ubuntu-0:3306", 
        "ssl": "DISABLED", 
        "status": "OK", 
        "statusText": "Cluster is ONLINE and can tolerate up to 2 failures.", 
        "topology": {
            "my-ubuntu-0:3306": {
                "address": "my-ubuntu-0:3306", 
                "memberRole": "PRIMARY", 
                "mode": "R/W", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-1:3306": {
                "address": "my-ubuntu-1:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-2:3306": {
                "address": "my-ubuntu-2:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-3:3306": {
                "address": "my-ubuntu-3:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-4:3306": {
                "address": "my-ubuntu-4:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }
        }, 
        "topologyMode": "Single-Primary"
    }, 
    "groupInformationSourceMember": "my-ubuntu-0:3306"
}

集群高可用性检查

要测试群集，我们可以使用样本码头组成的堆栈，该示例使用：

  • WordPress作为前端
  • mysqlrouter将将WP连接到MySQL

要运行此测试，您必须安装Docker和Docker compose。

用户和数据库创建

我们需要为WordPress创建一个DB和一个用户，为此，我们必须找到主服务器（检查群集状态并使用 - >“ mode”：“ r/w”）
找到节点。 

root@my-ubuntu-0:~# mysqlsh root@localhost
Please provide the password for 'root@localhost': ******************************************

MySQL  localhost:33060+ ssl  JS > \sql # <- SWITCH TO SQL MODE
Switching to SQL mode... Commands end with ;

create database wordpress;
create user 'wordpress'@'%' identified by 'wordpress';
grant all on wordpress.* TO 'wordpress'@'%';
flush privileges;

样品Dokcer组成堆栈

可以在examples文件夹中找到示例堆栈，这是组合文件：

version: '3.4'
services:
  wordpress:
    image: wordpress:latest
    ports:
      - 80:80
    restart: always
    environment:
      - WORDPRESS_DB_HOST=mysqlrouter:6446
      - WORDPRESS_DB_USER=wordpress
      - WORDPRESS_DB_PASSWORD=wordpress
      - WORDPRESS_DB_NAME=wordpress

  mysqlrouter:
    image: garutilorenzo/mysqlrouter:8.0.30
    volumes:
      - type: volume
        source: mysqlrouter
        target: /app/mysqlrouter/
        volume:
          nocopy: true
    environment:
     - MYSQL_HOST=my-ubuntu-0
     - MYSQL_PORT=3306
     - MYSQL_USER=root
     - MYSQL_PASSWORD=<CHANGE_ME> # <- the same password in the vars.yml file
     - MYSQL_ROUTER_ACCOUNT=mysql_router_user
     - MYSQL_ROUTER_PASSWORD=<CHANGE_ME> # <- openssl rand -base64 32 | sed 's/=//'
    extra_hosts:
      my-ubuntu-0: 192.168.25.110
      my-ubuntu-1: 192.168.25.111
      my-ubuntu-2: 192.168.25.112
      my-ubuntu-3: 192.168.25.113
      my-ubuntu-4: 192.168.25.114

volumes:
 mysqlrouter:

现在我们可以启动堆栈并检查日志：

docker-compose -f mysql-router-compose.yml up -d
docker-compose -f mysql-router-compose.yml logs mysqlrouter

examples-mysqlrouter-1  | Succesfully contacted mysql server at my-ubuntu-0. Checking for cluster state.
examples-mysqlrouter-1  | Check if config exist
examples-mysqlrouter-1  | bootstrap mysqlrouter with account mysql_router_user
examples-mysqlrouter-1  | Succesfully contacted mysql server at my-ubuntu-0. Trying to bootstrap.
examples-mysqlrouter-1  | Please enter MySQL password for root: 
examples-mysqlrouter-1  | # Bootstrapping MySQL Router instance at '/app/mysqlrouter'...
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | Please enter MySQL password for mysql_router_user: 
examples-mysqlrouter-1  | - Creating account(s) (only those that are needed, if any)
examples-mysqlrouter-1  | - Verifying account (using it to run SQL queries that would be run by Router)
examples-mysqlrouter-1  | - Storing account in keyring
examples-mysqlrouter-1  | - Adjusting permissions of generated files
examples-mysqlrouter-1  | - Creating configuration /app/mysqlrouter/mysqlrouter.conf
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | # MySQL Router configured for the InnoDB Cluster 'cluster_lab'
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | After this MySQL Router has been started with the generated configuration
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  |     $ mysqlrouter -c /app/mysqlrouter/mysqlrouter.conf
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | InnoDB Cluster 'cluster_lab' can be reached by connecting to:
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | ## MySQL Classic protocol
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | - Read/Write Connections: localhost:6446
examples-mysqlrouter-1  | - Read/Only Connections:  localhost:6447
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | ## MySQL X protocol
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | - Read/Write Connections: localhost:6448
examples-mysqlrouter-1  | - Read/Only Connections:  localhost:6449
examples-mysqlrouter-1  | 
examples-mysqlrouter-1  | Starting mysql-router.
examples-mysqlrouter-1  | 2022-08-30 12:03:57 io INFO [7f794f1e0bc0] starting 4 io-threads, using backend 'linux_epoll'
examples-mysqlrouter-1  | 2022-08-30 12:03:57 http_server INFO [7f794f1e0bc0] listening on 0.0.0.0:8443
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache_plugin INFO [7f794a606700] Starting Metadata Cache
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f794a606700] Connections using ssl_mode 'PREFERRED'
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700] Starting metadata cache refresh thread
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790e7fc700] [routing:bootstrap_rw] started: routing strategy = first-available
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790e7fc700] Start accepting connections for routing routing:bootstrap_rw listening on 6446
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790dffb700] [routing:bootstrap_x_ro] started: routing strategy = round-robin-with-fallback
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790dffb700] Start accepting connections for routing routing:bootstrap_x_ro listening on 6449
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790effd700] [routing:bootstrap_ro] started: routing strategy = round-robin-with-fallback
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790effd700] Start accepting connections for routing routing:bootstrap_ro listening on 6447
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700] Connected with metadata server running on my-ubuntu-2:3306
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790d7fa700] [routing:bootstrap_x_rw] started: routing strategy = first-available
examples-mysqlrouter-1  | 2022-08-30 12:03:57 routing INFO [7f790d7fa700] Start accepting connections for routing routing:bootstrap_x_rw listening on 6448
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700] Potential changes detected in cluster 'cluster_lab' after metadata refresh
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700] Metadata for cluster 'cluster_lab' has 5 member(s), single-primary: (view_id=0)
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700]     my-ubuntu-2:3306 / 33060 - mode=RO 
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700]     my-ubuntu-1:3306 / 33060 - mode=RO 
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700]     my-ubuntu-0:3306 / 33060 - mode=RW 
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700]     my-ubuntu-3:3306 / 33060 - mode=RO 
examples-mysqlrouter-1  | 2022-08-30 12:03:57 metadata_cache INFO [7f7948602700]     my-ubuntu-4:3306 / 33060 - mode=RO

测试前端

现在，如果您尝试访问localhost，您可以看到WordPress安装页面：

wp-install

安装和配置WP，现在我们已经准备好了一些Chaos Monkey

模拟灾难

要测试WP达到性能，我们可以开始此简单测试：

while true; do curl -s -o /dev/null -w "%{http_code}" http://localhost; echo; sleep 1; done
200
200

现在关闭RW节点（在这种情况下为my-ubuntu-0）：

root@my-ubuntu-0:~# sudo halt -p
Connection to 192.168.25.110 closed by remote host.
Connection to 192.168.25.110 closed.

并检查测试脚本的输出：

while true; do curl -s -o /dev/null -w "%{http_code}" http://localhost; echo; sleep 1; done
200
500 # <- my-ubuntu-0 shutdown and MySQL primary switch
200
200

现在，我们从第二个节点中检查群集状态，我们看到群集仍然是在线，并且可以忍受一次失败：

root@my-ubuntu-1:~# mysqlsh root@localhost
Please provide the password for 'root@localhost': ******************************************

MySQL  localhost:33060+ ssl  JS > clu = dba.getCluster()
MySQL  localhost:33060+ ssl  JS > clu.status()
{
    "clusterName": "cluster_lab", 
    "defaultReplicaSet": {
        "name": "default", 
        "primary": "my-ubuntu-2:3306", 
        "ssl": "DISABLED", 
        "status": "OK_PARTIAL", 
        "statusText": "Cluster is ONLINE and can tolerate up to ONE failure. 1 member is not active.", 
        "topology": {
            "my-ubuntu-0:3306": {
                "address": "my-ubuntu-0:3306", 
                "memberRole": "SECONDARY", 
                "mode": "n/a", 
                "readReplicas": {}, 
                "role": "HA", 
                "shellConnectError": "MySQL Error 2003: Could not open connection to 'my-ubuntu-0:3306': Can't connect to MySQL server on 'my-ubuntu-0:3306' (110)", 
                "status": "(MISSING)"
            }, 
            "my-ubuntu-1:3306": {
                "address": "my-ubuntu-1:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-2:3306": {
                "address": "my-ubuntu-2:3306", 
                "memberRole": "PRIMARY", 
                "mode": "R/W", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-3:3306": {
                "address": "my-ubuntu-3:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-4:3306": {
                "address": "my-ubuntu-4:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }
        }, 
        "topologyMode": "Single-Primary"
    }, 
    "groupInformationSourceMember": "my-ubuntu-2:3306"
}
 MySQL  localhost:33060+ ssl  JS >

如果您在localhost上检查WP，现在主人是 my-ubuntu-2 node。
现在，如果我们再次提出 my-ubuntu-0 节点，节点将重新加入群集，并将从其他节点获得更新：

root@my-ubuntu-1:~# mysqlsh root@localhost
Please provide the password for 'root@localhost': ******************************************

MySQL  localhost:33060+ ssl  JS > clu = dba.getCluster()
MySQL  localhost:33060+ ssl  JS > clu.status()
{
    "clusterName": "cluster_lab", 
    "defaultReplicaSet": {
        "name": "default", 
        "primary": "my-ubuntu-2:3306", 
        "ssl": "DISABLED", 
        "status": "OK", 
        "statusText": "Cluster is ONLINE and can tolerate up to 2 failures.", 
        "topology": {
            "my-ubuntu-0:3306": {
                "address": "my-ubuntu-0:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-1:3306": {
                "address": "my-ubuntu-1:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-2:3306": {
                "address": "my-ubuntu-2:3306", 
                "memberRole": "PRIMARY", 
                "mode": "R/W", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-3:3306": {
                "address": "my-ubuntu-3:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-4:3306": {
                "address": "my-ubuntu-4:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }
        }, 
        "topologyMode": "Single-Primary"
    }, 
    "groupInformationSourceMember": "my-ubuntu-2:3306"
}

从完全中断恢复

如果出于任何原因所有服务器都会下降，则必须从complete outage恢复集群。

要这样做，我们必须连接到一个实例，edit/etc/mysql/mysql.conf.d/innodb_cluster.cnf和set group_replication_bootstrap_group_group_group to on 和注释 group_replication_group_seeds 。我们现在必须重新启动mysql：

vagrant@my-ubuntu-0:~$
vi /etc/mysql/mysql.conf.d/innodb_cluster.cnf

group_replication_bootstrap_group=on
#group_replication_group_seeds=my-ubuntu-0:33061,my-ubuntu-1:33061,my-ubuntu-2:33061,my-ubuntu-3:33061,my-ubuntu-4:33061

systemctl restart mysqld

对于所有其他（四个）成员，我们必须启动组复制：

vagrant@my-ubuntu-4:~$ mysqlsh root@localhost
Please provide the password for 'root@localhost': ******************************************

MySQL  localhost:33060+ ssl  JS > \sql # <- SWITCH TO SQL MODE
Switching to SQL mode... Commands end with ;
MySQL  localhost:33060+ ssl  SQL > start group_replication;
Query OK, 0 rows affected (1.7095 sec)

如果集群上的流量很低或缺乏群集将在线：

MySQL  localhost:33060+ ssl  SQL > \js
Switching to JavaScript mode...
MySQL  localhost:33060+ ssl  JS > clu = dba.getCluster()
MySQL  localhost:33060+ ssl  JS > clu.status()
{
    "clusterName": "cluster_lab", 
    "defaultReplicaSet": {
        "name": "default", 
        "primary": "my-ubuntu-0:3306", 
        "ssl": "DISABLED", 
        "status": "OK", 
        "statusText": "Cluster is ONLINE and can tolerate up to 2 failures.", 
        "topology": {
            "my-ubuntu-0:3306": {
                "address": "my-ubuntu-0:3306", 
                "memberRole": "PRIMARY", 
                "mode": "R/W", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-1:3306": {
                "address": "my-ubuntu-1:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-2:3306": {
                "address": "my-ubuntu-2:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-3:3306": {
                "address": "my-ubuntu-3:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }, 
            "my-ubuntu-4:3306": {
                "address": "my-ubuntu-4:3306", 
                "memberRole": "SECONDARY", 
                "mode": "R/O", 
                "readReplicas": {}, 
                "replicationLag": "applier_queue_applied", 
                "role": "HA", 
                "status": "ONLINE", 
                "version": "8.0.30"
            }
        }, 
        "topologyMode": "Single-Primary"
    }, 
    "groupInformationSourceMember": "my-ubuntu-0:3306"
}

如果该集群在complete outage的时刻有大量的流量。您可能必须运行mysqlsh：

MySQL  localhost:33060+ ssl  JS > var clu = dba.rebootClusterFromCompleteOutage();

清理

完成后，您最终可以用：
销毁集群 

vagrant destroy

    my-ubuntu-4: Are you sure you want to destroy the 'my-ubuntu-4' VM? [y/N] y
==> my-ubuntu-4: Forcing shutdown of VM...
==> my-ubuntu-4: Destroying VM and associated drives...
    my-ubuntu-3: Are you sure you want to destroy the 'my-ubuntu-3' VM? [y/N] y
==> my-ubuntu-3: Forcing shutdown of VM...
==> my-ubuntu-3: Destroying VM and associated drives...
    my-ubuntu-2: Are you sure you want to destroy the 'my-ubuntu-2' VM? [y/N] y
==> my-ubuntu-2: Forcing shutdown of VM...
==> my-ubuntu-2: Destroying VM and associated drives...
    my-ubuntu-1: Are you sure you want to destroy the 'my-ubuntu-1' VM? [y/N] y
==> my-ubuntu-1: Forcing shutdown of VM...
==> my-ubuntu-1: Destroying VM and associated drives...
    my-ubuntu-0: Are you sure you want to destroy the 'my-ubuntu-0' VM? [y/N] y
==> my-ubuntu-0: Forcing shutdown of VM...
==> my-ubuntu-0: Destroying VM and associated drives...