mysql 8.0.18 mgr 搭建及其切换

mysql 8.0.18 mgr 搭建及其切换

转自:https://www.cnblogs.com/bjx2020/p/11810747.html

简介

集群自身不提供VIP机制,也没有像MongoDB副本集那样提供JAVA/PHP客户端API接口实现故障切换(需要开发自己写,成本较高),需要结合第三方HaProxy软件(建议2块网卡做bond0)+自定义脚本实现秒级故障切换,另通过代理方式,性能会降低,因为多了一层网络转发。

MySQL Group Replication有两种模式,单主模式single-primary mode和多主模式multi-primary mode,在同一个group内,不允许两种模式同时存在,并且若要切换到不同模式,必须修改配置后重新启动集群。

单主模式:

在单主模式下,只有一个节点可以可以读写,其他节点只能提供读,在单主模式下,该参数 group_replication_enforce_update_everywhere_checks 必须被设置为 FALSE ,当主节点宕掉,自动会根据服务器的server_uuid变量和group_replication_member_weight变量值,选择下一个slave谁作为主节点,group_replication_member_weight的值最高的成员被选为新的主节点,在group_replication_member_weight值相同的情况下,group根据数据字典中 server_uuid排序,排序在最前的被选择为主节点。

多主模式:

在多主模式下,在加入该群组的所有成员,所有服务器都设置为读写模式,在多主模式下,不支持SERIALIZABLE事务隔离级别,在多主模式下,不能完全支持级联外键约束。

Group Replication的特性和注意事项:

  • 全同步复制,事务要么在所有节点都提交,要么都回滚;
  • 多主复制,可以在任意节点进行写操作;
  • 在从服务器上并行应用事件,真正意义上的并行复制;
  • 节点自动配置–故障节点自动从集群中移除,当故障节点再次加入集群,无需手工备份当前数据库并拷贝至故障节点;
  • 应用程序的兼容性:无需更改应用程序,原生的MySQL接口;
  • 生产环境上集群推荐配置3个节点; 每个节点都包含完整的数据副本;
  • 各个节点的同步复制,通过GTID binlog ROW实现。

局限性:

  • 目前的复制仅仅支持InnoDB存储引擎;
  • 每张表必须有主键;
  • 只支持ipv4网络;
  • 集群最大支持9个节点;
  • 不支持Savepoints;
  • 不支持SERIALIZABLE隔离级别在Multi-Primary多主模式;
  • 不支持外键在Multi-Primary多主模式;
  • 整个集群的写入吞吐量是由最弱的节点限制,如果有一个节点变得缓慢,比如硬盘故障(RAID10坏了一块盘),那么整个集群将是缓慢的。为了稳定的高性能要求,所有的节点应使用统一的硬件。

优点:

  • 真正的多主架构,任何节点都可以进行读写,无需进行读写分离;(注:生产环境建议只在一台机器上写,由于集群是乐观锁并发控制,事务冲突的情况会在commit阶段发生。如果有两个事务在集群中不同的节点上对同一行写入并提交,失败的节点将回滚,客户端返回报错,作为DBA你不想被一群开发投诉的话,还是默默的开启Single-Primary写入模式);
  • 无集中管理,可以在任何时间点失去任何一个节点,集群将正常工作不受影响;
  • 节点宕机不会导致数据丢失;
  • 对应用透明。

缺点:

  • 加入新节点,开销大,需要复制完整的数据;
  • 不能有效的解决写扩展问题,磁盘空间满了,无法自动扩容,不能像MongoDB分片那样自动移动chunk做balance;
  • 有多少个节点就有多少份重复的数据,数据冗余;
  • 由于事务提交需要跨节点通讯(分布式事务),写入会比主从复制慢;
  • 对网络要求非常高,如果网络出现波动或机房被ARP攻击,造成两个节点失联,Group Replication集群发生脑裂,服务将不可用。

 

 

一、系统安装包

yum -y install make gcc-c++ cmake bison-devel ncurses-devel  readline-devel  libaio-devel perl libaio wget lrzsz vim libnuma* bzip2 xz

二、关闭防火墙和selinux

sed -i 's/SELINUX=enforcing/SELINUX=disabled/' /etc/selinux/config
setenforce  0

/etc/init.d/iptables stop

echo "/etc/init.d/iptables stop">>/etc/rc.local

三、修改系统限制参数

复制代码
复制代码
cat >> /etc/security/limits.conf << EOF
#
###custom
#
*           soft   nofile       20480
*           hard   nofile       65535
*           soft   nproc        20480
*           hard   nproc        65535
EOF
复制代码
复制代码

四、配置每台hosts主机解析

复制代码
复制代码
cat >> /etc/hosts <<"EOF"

10.10.146.28    bj-db-m1  
10.10.1.139     bj-db-m2  
10.10.173.84    bj-db-m3  

EOF
复制代码
复制代码

五、修改内核参数

复制代码
复制代码
cat >>/etc/sysctl.conf <<"EOF"
vm.swappiness=0
#增加tcp支持的队列数
net.ipv4.tcp_max_syn_backlog = 65535
#减少断开连接时 ,资源回收
net.ipv4.tcp_max_tw_buckets = 8000
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_tw_recycle = 1
net.ipv4.tcp_fin_timeout = 10
#改变本地的端口范围
net.ipv4.ip_local_port_range = 1024 65535
#允许更多的连接进入队列
net.ipv4.tcp_max_syn_backlog = 4096  
#对于只在本地使用的数据库服务器
net.ipv4.tcp_fin_timeout = 30
#端口监听队列
net.core.somaxconn=65535
#接受数据的速率
net.core.netdev_max_backlog=65535
net.core.wmem_default=87380
net.core.wmem_max=16777216
net.core.rmem_default=87380
net.core.rmem_max=16777216
EOF

sysctl -p
复制代码
复制代码

六、下载安装包

复制代码
复制代码
wget https://dev.mysql.com/get/Downloads/MySQL-8.0/mysql-8.0.18-linux-glibc2.12-x86_64.tar.xz
# 解压安装包
tar -xJf mysql-8.0.18-linux-glibc2.12-x86_64.tar.xz
# 进入目录,做软连接,方便以后升级
cd /usr/local/
ln -s /opt/mysql-8.0.18-linux-glibc2.12-x86_64 mysql
# 创建用户
groupadd mysql
useradd -g mysql mysql -d /home/mysql -s /sbin/nologin

# 创建相应的目录

mkdir -p /data/mysql/mysql_3306/{logs,tmp,undolog}
复制代码
复制代码

七、创建my.cnf配置文件

7-1、第一台配置

复制代码

# 第一台
if [ -f /etc/my.cnf ]; then
mv /etc/my.cnf /etc/my.cnf.`date +%Y%m%d%H%m`.bak
fi
# node1
cat >/data/mysql/mysql_3306/my_3306.cnf <<"EOF"
[client]
port = 3306
socket = /data/mysql/mysql_3306/tmp/mysql_3306.sock

[mysql]
prompt="\u@\h \R:\m:\s [\d]> "
no-auto-rehash

[mysqld]
user = mysql
port = 3306
admin_address = 127.0.0.1
basedir = /usr/local/mysql
datadir = /data/mysql/mysql_3306/data
socket = /data/mysql/mysql_3306/tmp/mysql_3306.sock
pid-file = mysql_3306.pid
character-set-server = utf8mb4
skip_name_resolve = 1

#replicate-wild-ignore-table=mysql.%
#replicate-wild-ignore-table=test.%
#replicate-wild-ignore-table=information_schema.%

# Two-Master configure
#server-1
#auto-increment-offset = 1
#auto-increment-increment = 2

#server-2
#auto-increment-offset = 2
#auto-increment-increment = 2


# semi sync replication settings #
#plugin_dir = /usr/local/mysql/lib/mysql/plugin
#plugin_load = "validate_password.so;rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
plugin_dir = /usr/local/mysql/lib/plugin #官方版本的路径
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so" #官方版本的路径

slave_parallel_workers = 4
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order = 1

open_files_limit = 65535
back_log = 1024
max_connections = 1024
max_connect_errors = 1000000
table_open_cache = 1024
table_definition_cache = 1024
table_open_cache_instances = 64
thread_stack = 512K
external-locking = FALSE
max_allowed_packet = 32M
sort_buffer_size = 4M
join_buffer_size = 4M
thread_cache_size = 1536
interactive_timeout = 600
wait_timeout = 600
tmp_table_size = 32M
max_heap_table_size = 32M
slow_query_log = 1
log_timestamps = SYSTEM
slow_query_log_file = /data/mysql/mysql_3306/logs/slow.log
log-error = /data/mysql/mysql_3306/logs/error.log
long_query_time = 0.1
log_queries_not_using_indexes =1
log_throttle_queries_not_using_indexes = 60
min_examined_row_limit = 100
log_slow_admin_statements = 1
log_slow_slave_statements = 1
server-id = 1423306
log-bin = /data/mysql/mysql_3306/logs/mysql-bin
sync_binlog = 1
binlog_cache_size = 4M
max_binlog_cache_size = 2G
max_binlog_size = 1G
binlog_expire_logs_seconds=2592000
master_info_repository = TABLE
relay_log_info_repository = TABLE
gtid_mode = on
enforce_gtid_consistency = 1
binlog_checksum=NONE
log_slave_updates
slave-rows-search-algorithms = 'INDEX_SCAN,HASH_SCAN'
binlog_format = row
binlog_row_image=FULL

relay_log_recovery = 1
relay-log-purge = 1
key_buffer_size = 32M
read_buffer_size = 8M
read_rnd_buffer_size = 4M
bulk_insert_buffer_size = 64M
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
lock_wait_timeout = 3600
explicit_defaults_for_timestamp = 1
innodb_thread_concurrency = 0
innodb_sync_spin_loops = 100
innodb_spin_wait_delay = 30

#transaction_isolation = REPEATABLE-READ
transaction_isolation = READ-COMMITTED
#innodb_additional_mem_pool_size = 16M
innodb_buffer_pool_size = 2867M
innodb_buffer_pool_instances = 4
innodb_buffer_pool_load_at_startup = 1
innodb_buffer_pool_dump_at_shutdown = 1
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 32M
innodb_log_file_size = 2G
innodb_log_files_in_group = 3
innodb_max_undo_log_size = 4G
innodb_undo_directory = /data/mysql/mysql_3306/undolog

# 根据您的服务器IOPS能力适当调整
# 一般配普通SSD盘的话,可以调整到 10000 - 20000
# 配置高端PCIe SSD卡的话,则可以调整的更高,比如 50000 - 80000
innodb_io_capacity = 4000
innodb_io_capacity_max = 8000
innodb_flush_sync = 0
innodb_flush_neighbors = 0
innodb_write_io_threads = 8
innodb_read_io_threads = 8
innodb_purge_threads = 4
innodb_page_cleaners = 4
innodb_open_files = 65535
innodb_max_dirty_pages_pct = 50
innodb_flush_method = O_DIRECT
innodb_lru_scan_depth = 4000
innodb_checksum_algorithm = crc32
innodb_lock_wait_timeout = 10
innodb_rollback_on_timeout = 1
innodb_print_all_deadlocks = 1
innodb_file_per_table = 1
innodb_online_alter_log_max_size = 4G
innodb_stats_on_metadata = 0

# some var for MySQL 8
log_error_verbosity = 3
innodb_print_ddl_logs = 1
binlog_expire_logs_seconds = 2592000
#innodb_dedicated_server = 0

innodb_status_file = 1
# 注意: 开启 innodb_status_output & innodb_status_output_locks 后, 可能会导致log-error文件增长较快
innodb_status_output = 0
innodb_status_output_locks = 0

#performance_schema
performance_schema = 1
performance_schema_instrument = '%memory%=on'
performance_schema_instrument = '%lock%=on'

#innodb monitor
innodb_monitor_enable="module_innodb"
innodb_monitor_enable="module_server"
innodb_monitor_enable="module_dml"
innodb_monitor_enable="module_ddl"
innodb_monitor_enable="module_trx"
innodb_monitor_enable="module_os"
innodb_monitor_enable="module_purge"
innodb_monitor_enable="module_log"
innodb_monitor_enable="module_lock"
innodb_monitor_enable="module_buffer"
innodb_monitor_enable="module_index"
innodb_monitor_enable="module_ibuf_system"
innodb_monitor_enable="module_buffer_page"
innodb_monitor_enable="module_adaptive_hash"

#MGR
#GR配置项 基中loose前缀表示若group Replication plugin未加载 mysql server仍明治维新启动
transaction_write_set_extraction = XXHASH64 #对每个事务获取write set,并且用XXHASH64算法获取hash值
loose-group_replication_group_name = "58f6e65e-9309-11e9-9d88-525400184a0a" #组名,此处可拿select uuid()生成
loose-group_replication_start_on_boot = off #在mysqld启动时不自动启动组复制
loose-group_replication_local_address = "10.10.146.28:33006" #本节点的IP地址和端口,注意该端口是组内成员之间通信的端口,而不是MySQL对外提供服务的端口
loose-group_replication_group_seeds = "10.10.146.28:33006,10.10.1.139:33006,10.10.173.84:33006" #种子节点的IP和端口号,新成员加入到集群的时候需要联系种子节点,启动集群的节点不使用该选项
loose-group_replication_bootstrap_group = off #关闭,如果打开会造成脑裂 #是否启动集群,注意,该选项任何时候只能用于一个节点,通常情况下启动集群的时候使用,启动之后需要关闭该选项

loose-group_replication_member_weight = 50 #权重选择


[mysqldump]
quick
max_allowed_packet = 32M

[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so
nice=-19
open-files-limit=65535

EOF


复制代码
# 第一台
if [ -f /etc/my.cnf ]; then
    mv /etc/my.cnf /etc/my.cnf.`date +%Y%m%d%H%m`.bak
fi 
# node1
cat >/data/mysql/mysql_3306/my_3306.cnf <<"EOF"
[client]
port    = 3306
socket    = /data/mysql/mysql_3306/tmp/mysql_3306.sock

[mysql]
prompt="\u@\h \R:\m:\s [\d]> "
no-auto-rehash

[mysqld]
user    = mysql
port    = 3306
admin_address = 127.0.0.1
basedir    = /usr/local/mysql
datadir    = /data/mysql/mysql_3306/data
socket    = /data/mysql/mysql_3306/tmp/mysql_3306.sock
pid-file = mysql_3306.pid
character-set-server = utf8mb4
skip_name_resolve = 1

#replicate-wild-ignore-table=mysql.%
#replicate-wild-ignore-table=test.%
#replicate-wild-ignore-table=information_schema.%

# Two-Master configure
#server-1 
#auto-increment-offset = 1
#auto-increment-increment = 2 

#server-2                          
#auto-increment-offset = 2
#auto-increment-increment = 2


# semi sync replication settings #
#plugin_dir = /usr/local/mysql/lib/mysql/plugin
#plugin_load = "validate_password.so;rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
plugin_dir = /usr/local/mysql/lib/plugin #官方版本的路径
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so" #官方版本的路径

slave_parallel_workers = 4
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order = 1

open_files_limit    = 65535
back_log = 1024
max_connections = 1024
max_connect_errors = 1000000
table_open_cache = 1024
table_definition_cache = 1024
table_open_cache_instances = 64
thread_stack = 512K
external-locking = FALSE
max_allowed_packet = 32M
sort_buffer_size = 4M
join_buffer_size = 4M
thread_cache_size = 1536
interactive_timeout = 600
wait_timeout = 600
tmp_table_size = 32M
max_heap_table_size = 32M
slow_query_log = 1
log_timestamps = SYSTEM
slow_query_log_file = /data/mysql/mysql_3306/logs/slow.log
log-error = /data/mysql/mysql_3306/logs/error.log
long_query_time = 0.1
log_queries_not_using_indexes =1
log_throttle_queries_not_using_indexes = 60
min_examined_row_limit = 100
log_slow_admin_statements = 1
log_slow_slave_statements = 1
server-id = 1423306
log-bin = /data/mysql/mysql_3306/logs/mysql-bin
sync_binlog = 1
binlog_cache_size = 4M
max_binlog_cache_size = 2G
max_binlog_size = 1G
binlog_expire_logs_seconds=2592000 
master_info_repository = TABLE
relay_log_info_repository = TABLE
gtid_mode = on
enforce_gtid_consistency = 1
binlog_checksum=NONE
log_slave_updates
slave-rows-search-algorithms = 'INDEX_SCAN,HASH_SCAN'
binlog_format = row
binlog_row_image=FULL

relay_log_recovery = 1
relay-log-purge = 1
key_buffer_size = 32M
read_buffer_size = 8M
read_rnd_buffer_size = 4M
bulk_insert_buffer_size = 64M
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
lock_wait_timeout = 3600
explicit_defaults_for_timestamp = 1
innodb_thread_concurrency = 0
innodb_sync_spin_loops = 100
innodb_spin_wait_delay = 30

#transaction_isolation = REPEATABLE-READ
transaction_isolation = READ-COMMITTED
#innodb_additional_mem_pool_size = 16M
innodb_buffer_pool_size = 2867M
innodb_buffer_pool_instances = 4
innodb_buffer_pool_load_at_startup = 1
innodb_buffer_pool_dump_at_shutdown = 1
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 32M
innodb_log_file_size = 2G
innodb_log_files_in_group = 3
innodb_max_undo_log_size = 4G
innodb_undo_directory = /data/mysql/mysql_3306/undolog

# 根据您的服务器IOPS能力适当调整
# 一般配普通SSD盘的话,可以调整到 10000 - 20000
# 配置高端PCIe SSD卡的话,则可以调整的更高,比如 50000 - 80000
innodb_io_capacity = 4000
innodb_io_capacity_max = 8000
innodb_flush_sync = 0
innodb_flush_neighbors = 0
innodb_write_io_threads = 8
innodb_read_io_threads = 8
innodb_purge_threads = 4
innodb_page_cleaners = 4
innodb_open_files = 65535
innodb_max_dirty_pages_pct = 50
innodb_flush_method = O_DIRECT
innodb_lru_scan_depth = 4000
innodb_checksum_algorithm = crc32
innodb_lock_wait_timeout = 10
innodb_rollback_on_timeout = 1
innodb_print_all_deadlocks = 1
innodb_file_per_table = 1
innodb_online_alter_log_max_size = 4G
innodb_stats_on_metadata = 0

# some var for MySQL 8
log_error_verbosity = 3
innodb_print_ddl_logs = 1
binlog_expire_logs_seconds = 2592000
#innodb_dedicated_server = 0

innodb_status_file = 1
# 注意: 开启 innodb_status_output & innodb_status_output_locks 后, 可能会导致log-error文件增长较快
innodb_status_output = 0
innodb_status_output_locks = 0

#performance_schema
performance_schema = 1
performance_schema_instrument = '%memory%=on'
performance_schema_instrument = '%lock%=on'

#innodb monitor
innodb_monitor_enable="module_innodb"
innodb_monitor_enable="module_server"
innodb_monitor_enable="module_dml"
innodb_monitor_enable="module_ddl"
innodb_monitor_enable="module_trx"
innodb_monitor_enable="module_os"
innodb_monitor_enable="module_purge"
innodb_monitor_enable="module_log"
innodb_monitor_enable="module_lock"
innodb_monitor_enable="module_buffer"
innodb_monitor_enable="module_index"
innodb_monitor_enable="module_ibuf_system"
innodb_monitor_enable="module_buffer_page"
innodb_monitor_enable="module_adaptive_hash"

#MGR
#GR配置项 基中loose前缀表示若group Replication plugin未加载 mysql server仍明治维新启动
transaction_write_set_extraction = XXHASH64 #对每个事务获取write set,并且用XXHASH64算法获取hash值
loose-group_replication_group_name = "58f6e65e-9309-11e9-9d88-525400184a0a" #组名,此处可拿select uuid()生成
loose-group_replication_start_on_boot = off #在mysqld启动时不自动启动组复制
loose-group_replication_local_address = "10.10.146.28:33006" #本节点的IP地址和端口,注意该端口是组内成员之间通信的端口,而不是MySQL对外提供服务的端口
loose-group_replication_group_seeds = "10.10.146.28:33006,10.10.1.139:33006,10.10.173.84:33006" #种子节点的IP和端口号,新成员加入到集群的时候需要联系种子节点,启动集群的节点不使用该选项
loose-group_replication_bootstrap_group = off #关闭,如果打开会造成脑裂 #是否启动集群,注意,该选项任何时候只能用于一个节点,通常情况下启动集群的时候使用,启动之后需要关闭该选项

loose-group_replication_member_weight = 50 #权重选择


[mysqldump]
quick
max_allowed_packet = 32M

[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so 
nice=-19
open-files-limit=65535

EOF
复制代码
复制代码

7-2、第二台配置

复制代码

# 第二台
if [ -f /etc/my.cnf ]; then
mv /etc/my.cnf /etc/my.cnf.`date +%Y%m%d%H%m`.bak
fi
# node1
cat >/data/mysql/mysql_3306/my_3306.cnf <<"EOF"
[client]
port = 3306
socket = /data/mysql/mysql_3306/tmp/mysql_3306.sock

[mysql]
prompt="\u@\h \R:\m:\s [\d]> "
no-auto-rehash

[mysqld]
user = mysql
port = 3306
admin_address = 127.0.0.1
basedir = /usr/local/mysql
datadir = /data/mysql/mysql_3306/data
socket = /data/mysql/mysql_3306/tmp/mysql_3306.sock
pid-file = mysql_3306.pid
character-set-server = utf8mb4
skip_name_resolve = 1

#replicate-wild-ignore-table=mysql.%
#replicate-wild-ignore-table=test.%
#replicate-wild-ignore-table=information_schema.%

# Two-Master configure
#server-1
#auto-increment-offset = 1
#auto-increment-increment = 2

#server-2
#auto-increment-offset = 2
#auto-increment-increment = 2


# semi sync replication settings #
#plugin_dir = /usr/local/mysql/lib/mysql/plugin
#plugin_load = "validate_password.so;rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
plugin_dir = /usr/local/mysql/lib/plugin #官方版本的路径
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so" #官方版本的路径

slave_parallel_workers = 4
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order = 1

open_files_limit = 65535
back_log = 1024
max_connections = 1024
max_connect_errors = 1000000
table_open_cache = 1024
table_definition_cache = 1024
table_open_cache_instances = 64
thread_stack = 512K
external-locking = FALSE
max_allowed_packet = 32M
sort_buffer_size = 4M
join_buffer_size = 4M
thread_cache_size = 1536
interactive_timeout = 600
wait_timeout = 600
tmp_table_size = 32M
max_heap_table_size = 32M
slow_query_log = 1
log_timestamps = SYSTEM
slow_query_log_file = /data/mysql/mysql_3306/logs/slow.log
log-error = /data/mysql/mysql_3306/logs/error.log
long_query_time = 0.1
log_queries_not_using_indexes =1
log_throttle_queries_not_using_indexes = 60
min_examined_row_limit = 100
log_slow_admin_statements = 1
log_slow_slave_statements = 1
server-id = 1433306
log-bin = /data/mysql/mysql_3306/logs/mysql-bin
sync_binlog = 1
binlog_cache_size = 4M
max_binlog_cache_size = 2G
max_binlog_size = 1G
binlog_expire_logs_seconds=2592000
master_info_repository = TABLE
relay_log_info_repository = TABLE
gtid_mode = on
enforce_gtid_consistency = 1
binlog_checksum=NONE
log_slave_updates
slave-rows-search-algorithms = 'INDEX_SCAN,HASH_SCAN'
binlog_format = row
binlog_row_image=FULL

relay_log_recovery = 1
relay-log-purge = 1
key_buffer_size = 32M
read_buffer_size = 8M
read_rnd_buffer_size = 4M
bulk_insert_buffer_size = 64M
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
lock_wait_timeout = 3600
explicit_defaults_for_timestamp = 1
innodb_thread_concurrency = 0
innodb_sync_spin_loops = 100
innodb_spin_wait_delay = 30

#transaction_isolation = REPEATABLE-READ
transaction_isolation = READ-COMMITTED
#innodb_additional_mem_pool_size = 16M
innodb_buffer_pool_size = 2867M
innodb_buffer_pool_instances = 4
innodb_buffer_pool_load_at_startup = 1
innodb_buffer_pool_dump_at_shutdown = 1
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 32M
innodb_log_file_size = 2G
innodb_log_files_in_group = 3
innodb_max_undo_log_size = 4G
innodb_undo_directory = /data/mysql/mysql_3306/undolog

# 根据您的服务器IOPS能力适当调整
# 一般配普通SSD盘的话,可以调整到 10000 - 20000
# 配置高端PCIe SSD卡的话,则可以调整的更高,比如 50000 - 80000
innodb_io_capacity = 4000
innodb_io_capacity_max = 8000
innodb_flush_sync = 0
innodb_flush_neighbors = 0
innodb_write_io_threads = 8
innodb_read_io_threads = 8
innodb_purge_threads = 4
innodb_page_cleaners = 4
innodb_open_files = 65535
innodb_max_dirty_pages_pct = 50
innodb_flush_method = O_DIRECT
innodb_lru_scan_depth = 4000
innodb_checksum_algorithm = crc32
innodb_lock_wait_timeout = 10
innodb_rollback_on_timeout = 1
innodb_print_all_deadlocks = 1
innodb_file_per_table = 1
innodb_online_alter_log_max_size = 4G
innodb_stats_on_metadata = 0

# some var for MySQL 8
log_error_verbosity = 3
innodb_print_ddl_logs = 1
binlog_expire_logs_seconds = 2592000
#innodb_dedicated_server = 0

innodb_status_file = 1
# 注意: 开启 innodb_status_output & innodb_status_output_locks 后, 可能会导致log-error文件增长较快
innodb_status_output = 0
innodb_status_output_locks = 0

#performance_schema
performance_schema = 1
performance_schema_instrument = '%memory%=on'
performance_schema_instrument = '%lock%=on'

#innodb monitor
innodb_monitor_enable="module_innodb"
innodb_monitor_enable="module_server"
innodb_monitor_enable="module_dml"
innodb_monitor_enable="module_ddl"
innodb_monitor_enable="module_trx"
innodb_monitor_enable="module_os"
innodb_monitor_enable="module_purge"
innodb_monitor_enable="module_log"
innodb_monitor_enable="module_lock"
innodb_monitor_enable="module_buffer"
innodb_monitor_enable="module_index"
innodb_monitor_enable="module_ibuf_system"
innodb_monitor_enable="module_buffer_page"
innodb_monitor_enable="module_adaptive_hash"

#MGR
#GR配置项 基中loose前缀表示若group Replication plugin未加载 mysql server仍明治维新启动
transaction_write_set_extraction = XXHASH64 #对每个事务获取write set,并且用XXHASH64算法获取hash值
loose-group_replication_group_name = "58f6e65e-9309-11e9-9d88-525400184a0a" #组名,此处可拿select uuid()生成
loose-group_replication_start_on_boot = off #在mysqld启动时不自动启动组复制
loose-group_replication_local_address = "10.10.1.139:33006" #本节点的IP地址和端口,注意该端口是组内成员之间通信的端口,而不是MySQL对外提供服务的端口
loose-group_replication_group_seeds = "10.10.146.28:33006,10.10.1.139:33006,10.10.173.84:33006" #种子节点的IP和端口号,新成员加入到集群的时候需要联系种子节点,启动集群的节点不使用该选项
loose-group_replication_bootstrap_group = off #关闭,如果打开会造成脑裂 #是否启动集群,注意,该选项任何时候只能用于一个节点,通常情况下启动集群的时候使用,启动之后需要关闭该选项

loose-group_replication_member_weight = 50 #权重选择


[mysqldump]
quick
max_allowed_packet = 32M

[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so
nice=-19
open-files-limit=65535

EOF


复制代码
# 第二台
if [ -f /etc/my.cnf ]; then
    mv /etc/my.cnf /etc/my.cnf.`date +%Y%m%d%H%m`.bak
fi 
# node1
cat >/data/mysql/mysql_3306/my_3306.cnf <<"EOF"
[client]
port    = 3306
socket    = /data/mysql/mysql_3306/tmp/mysql_3306.sock

[mysql]
prompt="\u@\h \R:\m:\s [\d]> "
no-auto-rehash

[mysqld]
user    = mysql
port    = 3306
admin_address = 127.0.0.1
basedir    = /usr/local/mysql
datadir    = /data/mysql/mysql_3306/data
socket    = /data/mysql/mysql_3306/tmp/mysql_3306.sock
pid-file = mysql_3306.pid
character-set-server = utf8mb4
skip_name_resolve = 1

#replicate-wild-ignore-table=mysql.%
#replicate-wild-ignore-table=test.%
#replicate-wild-ignore-table=information_schema.%

# Two-Master configure
#server-1 
#auto-increment-offset = 1
#auto-increment-increment = 2 

#server-2                          
#auto-increment-offset = 2
#auto-increment-increment = 2


# semi sync replication settings #
#plugin_dir = /usr/local/mysql/lib/mysql/plugin
#plugin_load = "validate_password.so;rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
plugin_dir = /usr/local/mysql/lib/plugin #官方版本的路径
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so" #官方版本的路径

slave_parallel_workers = 4
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order = 1

open_files_limit    = 65535
back_log = 1024
max_connections = 1024
max_connect_errors = 1000000
table_open_cache = 1024
table_definition_cache = 1024
table_open_cache_instances = 64
thread_stack = 512K
external-locking = FALSE
max_allowed_packet = 32M
sort_buffer_size = 4M
join_buffer_size = 4M
thread_cache_size = 1536
interactive_timeout = 600
wait_timeout = 600
tmp_table_size = 32M
max_heap_table_size = 32M
slow_query_log = 1
log_timestamps = SYSTEM
slow_query_log_file = /data/mysql/mysql_3306/logs/slow.log
log-error = /data/mysql/mysql_3306/logs/error.log
long_query_time = 0.1
log_queries_not_using_indexes =1
log_throttle_queries_not_using_indexes = 60
min_examined_row_limit = 100
log_slow_admin_statements = 1
log_slow_slave_statements = 1
server-id = 1433306
log-bin = /data/mysql/mysql_3306/logs/mysql-bin
sync_binlog = 1
binlog_cache_size = 4M
max_binlog_cache_size = 2G
max_binlog_size = 1G
binlog_expire_logs_seconds=2592000 
master_info_repository = TABLE
relay_log_info_repository = TABLE
gtid_mode = on
enforce_gtid_consistency = 1
binlog_checksum=NONE
log_slave_updates
slave-rows-search-algorithms = 'INDEX_SCAN,HASH_SCAN'
binlog_format = row
binlog_row_image=FULL

relay_log_recovery = 1
relay-log-purge = 1
key_buffer_size = 32M
read_buffer_size = 8M
read_rnd_buffer_size = 4M
bulk_insert_buffer_size = 64M
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
lock_wait_timeout = 3600
explicit_defaults_for_timestamp = 1
innodb_thread_concurrency = 0
innodb_sync_spin_loops = 100
innodb_spin_wait_delay = 30

#transaction_isolation = REPEATABLE-READ
transaction_isolation = READ-COMMITTED
#innodb_additional_mem_pool_size = 16M
innodb_buffer_pool_size = 2867M
innodb_buffer_pool_instances = 4
innodb_buffer_pool_load_at_startup = 1
innodb_buffer_pool_dump_at_shutdown = 1
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 32M
innodb_log_file_size = 2G
innodb_log_files_in_group = 3
innodb_max_undo_log_size = 4G
innodb_undo_directory = /data/mysql/mysql_3306/undolog

# 根据您的服务器IOPS能力适当调整
# 一般配普通SSD盘的话,可以调整到 10000 - 20000
# 配置高端PCIe SSD卡的话,则可以调整的更高,比如 50000 - 80000
innodb_io_capacity = 4000
innodb_io_capacity_max = 8000
innodb_flush_sync = 0
innodb_flush_neighbors = 0
innodb_write_io_threads = 8
innodb_read_io_threads = 8
innodb_purge_threads = 4
innodb_page_cleaners = 4
innodb_open_files = 65535
innodb_max_dirty_pages_pct = 50
innodb_flush_method = O_DIRECT
innodb_lru_scan_depth = 4000
innodb_checksum_algorithm = crc32
innodb_lock_wait_timeout = 10
innodb_rollback_on_timeout = 1
innodb_print_all_deadlocks = 1
innodb_file_per_table = 1
innodb_online_alter_log_max_size = 4G
innodb_stats_on_metadata = 0

# some var for MySQL 8
log_error_verbosity = 3
innodb_print_ddl_logs = 1
binlog_expire_logs_seconds = 2592000
#innodb_dedicated_server = 0

innodb_status_file = 1
# 注意: 开启 innodb_status_output & innodb_status_output_locks 后, 可能会导致log-error文件增长较快
innodb_status_output = 0
innodb_status_output_locks = 0

#performance_schema
performance_schema = 1
performance_schema_instrument = '%memory%=on'
performance_schema_instrument = '%lock%=on'

#innodb monitor
innodb_monitor_enable="module_innodb"
innodb_monitor_enable="module_server"
innodb_monitor_enable="module_dml"
innodb_monitor_enable="module_ddl"
innodb_monitor_enable="module_trx"
innodb_monitor_enable="module_os"
innodb_monitor_enable="module_purge"
innodb_monitor_enable="module_log"
innodb_monitor_enable="module_lock"
innodb_monitor_enable="module_buffer"
innodb_monitor_enable="module_index"
innodb_monitor_enable="module_ibuf_system"
innodb_monitor_enable="module_buffer_page"
innodb_monitor_enable="module_adaptive_hash"

#MGR
#GR配置项 基中loose前缀表示若group Replication plugin未加载 mysql server仍明治维新启动
transaction_write_set_extraction = XXHASH64 #对每个事务获取write set,并且用XXHASH64算法获取hash值
loose-group_replication_group_name = "58f6e65e-9309-11e9-9d88-525400184a0a" #组名,此处可拿select uuid()生成
loose-group_replication_start_on_boot = off #在mysqld启动时不自动启动组复制
loose-group_replication_local_address = "10.10.1.139:33006" #本节点的IP地址和端口,注意该端口是组内成员之间通信的端口,而不是MySQL对外提供服务的端口
loose-group_replication_group_seeds = "10.10.146.28:33006,10.10.1.139:33006,10.10.173.84:33006" #种子节点的IP和端口号,新成员加入到集群的时候需要联系种子节点,启动集群的节点不使用该选项
loose-group_replication_bootstrap_group = off #关闭,如果打开会造成脑裂 #是否启动集群,注意,该选项任何时候只能用于一个节点,通常情况下启动集群的时候使用,启动之后需要关闭该选项

loose-group_replication_member_weight = 50 #权重选择


[mysqldump]
quick
max_allowed_packet = 32M

[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so 
nice=-19
open-files-limit=65535

EOF
复制代码
复制代码

7-3、第三台配置

复制代码

# 第三台
if [ -f /etc/my.cnf ]; then
mv /etc/my.cnf /etc/my.cnf.`date +%Y%m%d%H%m`.bak
fi
# node1
cat >/data/mysql/mysql_3306/my_3306.cnf <<"EOF"
[client]
port = 3306
socket = /data/mysql/mysql_3306/tmp/mysql_3306.sock

[mysql]
prompt="\u@\h \R:\m:\s [\d]> "
no-auto-rehash

[mysqld]
user = mysql
port = 3306
admin_address = 127.0.0.1
basedir = /usr/local/mysql
datadir = /data/mysql/mysql_3306/data
socket = /data/mysql/mysql_3306/tmp/mysql_3306.sock
pid-file = mysql_3306.pid
character-set-server = utf8mb4
skip_name_resolve = 1

#replicate-wild-ignore-table=mysql.%
#replicate-wild-ignore-table=test.%
#replicate-wild-ignore-table=information_schema.%

# Two-Master configure
#server-1
#auto-increment-offset = 1
#auto-increment-increment = 2

#server-2
#auto-increment-offset = 2
#auto-increment-increment = 2


# semi sync replication settings #
#plugin_dir = /usr/local/mysql/lib/mysql/plugin
#plugin_load = "validate_password.so;rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
plugin_dir = /usr/local/mysql/lib/plugin #官方版本的路径
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so" #官方版本的路径

slave_parallel_workers = 4
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order = 1

open_files_limit = 65535
back_log = 1024
max_connections = 1024
max_connect_errors = 1000000
table_open_cache = 1024
table_definition_cache = 1024
table_open_cache_instances = 64
thread_stack = 512K
external-locking = FALSE
max_allowed_packet = 32M
sort_buffer_size = 4M
join_buffer_size = 4M
thread_cache_size = 1536
interactive_timeout = 600
wait_timeout = 600
tmp_table_size = 32M
max_heap_table_size = 32M
slow_query_log = 1
log_timestamps = SYSTEM
slow_query_log_file = /data/mysql/mysql_3306/logs/slow.log
log-error = /data/mysql/mysql_3306/logs/error.log
long_query_time = 0.1
log_queries_not_using_indexes =1
log_throttle_queries_not_using_indexes = 60
min_examined_row_limit = 100
log_slow_admin_statements = 1
log_slow_slave_statements = 1
server-id = 1443306
log-bin = /data/mysql/mysql_3306/logs/mysql-bin
sync_binlog = 1
binlog_cache_size = 4M
max_binlog_cache_size = 2G
max_binlog_size = 1G
binlog_expire_logs_seconds=2592000
master_info_repository = TABLE
relay_log_info_repository = TABLE
gtid_mode = on
enforce_gtid_consistency = 1
binlog_checksum=NONE
log_slave_updates
slave-rows-search-algorithms = 'INDEX_SCAN,HASH_SCAN'
binlog_format = row
binlog_row_image=FULL

relay_log_recovery = 1
relay-log-purge = 1
key_buffer_size = 32M
read_buffer_size = 8M
read_rnd_buffer_size = 4M
bulk_insert_buffer_size = 64M
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
lock_wait_timeout = 3600
explicit_defaults_for_timestamp = 1
innodb_thread_concurrency = 0
innodb_sync_spin_loops = 100
innodb_spin_wait_delay = 30

#transaction_isolation = REPEATABLE-READ
transaction_isolation = READ-COMMITTED
#innodb_additional_mem_pool_size = 16M
innodb_buffer_pool_size = 2867M
innodb_buffer_pool_instances = 4
innodb_buffer_pool_load_at_startup = 1
innodb_buffer_pool_dump_at_shutdown = 1
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 32M
innodb_log_file_size = 2G
innodb_log_files_in_group = 3
innodb_max_undo_log_size = 4G
innodb_undo_directory = /data/mysql/mysql_3306/undolog

# 根据您的服务器IOPS能力适当调整
# 一般配普通SSD盘的话,可以调整到 10000 - 20000
# 配置高端PCIe SSD卡的话,则可以调整的更高,比如 50000 - 80000
innodb_io_capacity = 4000
innodb_io_capacity_max = 8000
innodb_flush_sync = 0
innodb_flush_neighbors = 0
innodb_write_io_threads = 8
innodb_read_io_threads = 8
innodb_purge_threads = 4
innodb_page_cleaners = 4
innodb_open_files = 65535
innodb_max_dirty_pages_pct = 50
innodb_flush_method = O_DIRECT
innodb_lru_scan_depth = 4000
innodb_checksum_algorithm = crc32
innodb_lock_wait_timeout = 10
innodb_rollback_on_timeout = 1
innodb_print_all_deadlocks = 1
innodb_file_per_table = 1
innodb_online_alter_log_max_size = 4G
innodb_stats_on_metadata = 0

# some var for MySQL 8
log_error_verbosity = 3
innodb_print_ddl_logs = 1
binlog_expire_logs_seconds = 2592000
#innodb_dedicated_server = 0

innodb_status_file = 1
# 注意: 开启 innodb_status_output & innodb_status_output_locks 后, 可能会导致log-error文件增长较快
innodb_status_output = 0
innodb_status_output_locks = 0

#performance_schema
performance_schema = 1
performance_schema_instrument = '%memory%=on'
performance_schema_instrument = '%lock%=on'

#innodb monitor
innodb_monitor_enable="module_innodb"
innodb_monitor_enable="module_server"
innodb_monitor_enable="module_dml"
innodb_monitor_enable="module_ddl"
innodb_monitor_enable="module_trx"
innodb_monitor_enable="module_os"
innodb_monitor_enable="module_purge"
innodb_monitor_enable="module_log"
innodb_monitor_enable="module_lock"
innodb_monitor_enable="module_buffer"
innodb_monitor_enable="module_index"
innodb_monitor_enable="module_ibuf_system"
innodb_monitor_enable="module_buffer_page"
innodb_monitor_enable="module_adaptive_hash"

#MGR
#GR配置项 基中loose前缀表示若group Replication plugin未加载 mysql server仍明治维新启动
transaction_write_set_extraction = XXHASH64 #对每个事务获取write set,并且用XXHASH64算法获取hash值
loose-group_replication_group_name = "58f6e65e-9309-11e9-9d88-525400184a0a" #组名,此处可拿select uuid()生成
loose-group_replication_start_on_boot = off #在mysqld启动时不自动启动组复制
loose-group_replication_local_address = "10.10.173.84:33006" #本节点的IP地址和端口,注意该端口是组内成员之间通信的端口,而不是MySQL对外提供服务的端口
loose-group_replication_group_seeds = "10.10.146.28:33006,10.10.1.139:33006,10.10.173.84:33006" #种子节点的IP和端口号,新成员加入到集群的时候需要联系种子节点,启动集群的节点不使用该选项
loose-group_replication_bootstrap_group = off #关闭,如果打开会造成脑裂 #是否启动集群,注意,该选项任何时候只能用于一个节点,通常情况下启动集群的时候使用,启动之后需要关闭该选项

loose-group_replication_member_weight = 50 #权重选择


[mysqldump]
quick
max_allowed_packet = 32M

[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so
nice=-19
open-files-limit=65535

EOF


复制代码
# 第三台
if [ -f /etc/my.cnf ]; then
    mv /etc/my.cnf /etc/my.cnf.`date +%Y%m%d%H%m`.bak
fi 
# node1
cat >/data/mysql/mysql_3306/my_3306.cnf <<"EOF"
[client]
port    = 3306
socket    = /data/mysql/mysql_3306/tmp/mysql_3306.sock

[mysql]
prompt="\u@\h \R:\m:\s [\d]> "
no-auto-rehash

[mysqld]
user    = mysql
port    = 3306
admin_address = 127.0.0.1
basedir    = /usr/local/mysql
datadir    = /data/mysql/mysql_3306/data
socket    = /data/mysql/mysql_3306/tmp/mysql_3306.sock
pid-file = mysql_3306.pid
character-set-server = utf8mb4
skip_name_resolve = 1

#replicate-wild-ignore-table=mysql.%
#replicate-wild-ignore-table=test.%
#replicate-wild-ignore-table=information_schema.%

# Two-Master configure
#server-1 
#auto-increment-offset = 1
#auto-increment-increment = 2 

#server-2                          
#auto-increment-offset = 2
#auto-increment-increment = 2


# semi sync replication settings #
#plugin_dir = /usr/local/mysql/lib/mysql/plugin
#plugin_load = "validate_password.so;rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
plugin_dir = /usr/local/mysql/lib/plugin #官方版本的路径
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so" #官方版本的路径

slave_parallel_workers = 4
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order = 1

open_files_limit    = 65535
back_log = 1024
max_connections = 1024
max_connect_errors = 1000000
table_open_cache = 1024
table_definition_cache = 1024
table_open_cache_instances = 64
thread_stack = 512K
external-locking = FALSE
max_allowed_packet = 32M
sort_buffer_size = 4M
join_buffer_size = 4M
thread_cache_size = 1536
interactive_timeout = 600
wait_timeout = 600
tmp_table_size = 32M
max_heap_table_size = 32M
slow_query_log = 1
log_timestamps = SYSTEM
slow_query_log_file = /data/mysql/mysql_3306/logs/slow.log
log-error = /data/mysql/mysql_3306/logs/error.log
long_query_time = 0.1
log_queries_not_using_indexes =1
log_throttle_queries_not_using_indexes = 60
min_examined_row_limit = 100
log_slow_admin_statements = 1
log_slow_slave_statements = 1
server-id = 1443306
log-bin = /data/mysql/mysql_3306/logs/mysql-bin
sync_binlog = 1
binlog_cache_size = 4M
max_binlog_cache_size = 2G
max_binlog_size = 1G
binlog_expire_logs_seconds=2592000 
master_info_repository = TABLE
relay_log_info_repository = TABLE
gtid_mode = on
enforce_gtid_consistency = 1
binlog_checksum=NONE
log_slave_updates
slave-rows-search-algorithms = 'INDEX_SCAN,HASH_SCAN'
binlog_format = row
binlog_row_image=FULL

relay_log_recovery = 1
relay-log-purge = 1
key_buffer_size = 32M
read_buffer_size = 8M
read_rnd_buffer_size = 4M
bulk_insert_buffer_size = 64M
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
lock_wait_timeout = 3600
explicit_defaults_for_timestamp = 1
innodb_thread_concurrency = 0
innodb_sync_spin_loops = 100
innodb_spin_wait_delay = 30

#transaction_isolation = REPEATABLE-READ
transaction_isolation = READ-COMMITTED
#innodb_additional_mem_pool_size = 16M
innodb_buffer_pool_size = 2867M
innodb_buffer_pool_instances = 4
innodb_buffer_pool_load_at_startup = 1
innodb_buffer_pool_dump_at_shutdown = 1
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 32M
innodb_log_file_size = 2G
innodb_log_files_in_group = 3
innodb_max_undo_log_size = 4G
innodb_undo_directory = /data/mysql/mysql_3306/undolog

# 根据您的服务器IOPS能力适当调整
# 一般配普通SSD盘的话,可以调整到 10000 - 20000
# 配置高端PCIe SSD卡的话,则可以调整的更高,比如 50000 - 80000
innodb_io_capacity = 4000
innodb_io_capacity_max = 8000
innodb_flush_sync = 0
innodb_flush_neighbors = 0
innodb_write_io_threads = 8
innodb_read_io_threads = 8
innodb_purge_threads = 4
innodb_page_cleaners = 4
innodb_open_files = 65535
innodb_max_dirty_pages_pct = 50
innodb_flush_method = O_DIRECT
innodb_lru_scan_depth = 4000
innodb_checksum_algorithm = crc32
innodb_lock_wait_timeout = 10
innodb_rollback_on_timeout = 1
innodb_print_all_deadlocks = 1
innodb_file_per_table = 1
innodb_online_alter_log_max_size = 4G
innodb_stats_on_metadata = 0

# some var for MySQL 8
log_error_verbosity = 3
innodb_print_ddl_logs = 1
binlog_expire_logs_seconds = 2592000
#innodb_dedicated_server = 0

innodb_status_file = 1
# 注意: 开启 innodb_status_output & innodb_status_output_locks 后, 可能会导致log-error文件增长较快
innodb_status_output = 0
innodb_status_output_locks = 0

#performance_schema
performance_schema = 1
performance_schema_instrument = '%memory%=on'
performance_schema_instrument = '%lock%=on'

#innodb monitor
innodb_monitor_enable="module_innodb"
innodb_monitor_enable="module_server"
innodb_monitor_enable="module_dml"
innodb_monitor_enable="module_ddl"
innodb_monitor_enable="module_trx"
innodb_monitor_enable="module_os"
innodb_monitor_enable="module_purge"
innodb_monitor_enable="module_log"
innodb_monitor_enable="module_lock"
innodb_monitor_enable="module_buffer"
innodb_monitor_enable="module_index"
innodb_monitor_enable="module_ibuf_system"
innodb_monitor_enable="module_buffer_page"
innodb_monitor_enable="module_adaptive_hash"

#MGR
#GR配置项 基中loose前缀表示若group Replication plugin未加载 mysql server仍明治维新启动
transaction_write_set_extraction = XXHASH64 #对每个事务获取write set,并且用XXHASH64算法获取hash值
loose-group_replication_group_name = "58f6e65e-9309-11e9-9d88-525400184a0a" #组名,此处可拿select uuid()生成
loose-group_replication_start_on_boot = off #在mysqld启动时不自动启动组复制
loose-group_replication_local_address = "10.10.173.84:33006" #本节点的IP地址和端口,注意该端口是组内成员之间通信的端口,而不是MySQL对外提供服务的端口
loose-group_replication_group_seeds = "10.10.146.28:33006,10.10.1.139:33006,10.10.173.84:33006" #种子节点的IP和端口号,新成员加入到集群的时候需要联系种子节点,启动集群的节点不使用该选项
loose-group_replication_bootstrap_group = off #关闭,如果打开会造成脑裂 #是否启动集群,注意,该选项任何时候只能用于一个节点,通常情况下启动集群的时候使用,启动之后需要关闭该选项

loose-group_replication_member_weight = 50 #权重选择


[mysqldump]
quick
max_allowed_packet = 32M

[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so 
nice=-19
open-files-limit=65535

EOF
复制代码
复制代码

八、修改权限、初始化并启动

复制代码
复制代码
chown -R mysql.mysql /data/mysql/mysql_3306
chown -R mysql.mysql /usr/local/mysql/

#初始化
# /usr/local/mysql/bin/mysqld --user=mysql --basedir=/usr/local/mysql --datadir=/data/mysql --initialize-insecure
# 官方推荐使用--initialize,会在错误日志中生成难以输入的临时密码,我这里使用的免密码的方式。
/usr/local/mysql/bin/mysqld --defaults-file=/data/mysql/mysql_3306/my_3306.cnf --initialize-insecure  --user=mysql &

#启动数据库
/usr/local/mysql/bin/mysqld_safe --defaults-file=/data/mysql/mysql_3306/my_3306.cnf &
复制代码
复制代码

九、查看日志

#9、查看日志
# tail -f /data/mysql/mysql_3306/logs/error.log

十、初次登陆

#10、初次登陆
/usr/local/mysql/bin/mysql -S /data/mysql/mysql_3306/tmp/mysql_3306.sock

十一、修改密码

复制代码
复制代码
# 修改密码方法

set sql_log_bin = 0;

ALTER USER 'root'@'localhost' IDENTIFIED WITH mysql_native_password BY 'GJjumB6g4FcwdF3R6AZc' PASSWORD EXPIRE NEVER ;
create user 'root'@'127.0.0.1' identified WITH mysql_native_password by 'GJjumB6g4FcwdF3R6AZc' PASSWORD EXPIRE NEVER ;
grant all privileges on *.* to 'root'@'127.0.0.1' with grant option;

create user 'admin_m'@'127.0.0.1' identified WITH mysql_native_password by 'fcfmTbRw1tz2x5L5GvjJ' PASSWORD EXPIRE NEVER ; 
grant all privileges on *.* to 'admin_m'@'127.0.0.1' with grant option;

create user 'admin_m'@'%' identified WITH mysql_native_password by 'fcfmTbRw1tz2x5L5GvjJ' PASSWORD EXPIRE NEVER ; 
grant all privileges on *.* to 'admin_m'@'%' with grant option;

create user 'test_w'@'%' identified with mysql_native_password by 'EeCrfUDO6wRzn72BBQ52' PASSWORD EXPIRE NEVER ;
grant insert,delete,update,select on db144.* to 'test_w'@'%' ;

create user 'test_r'@'%' identified with mysql_native_password by 'EeCrfUDO6wRzn72BBQ52' PASSWORD EXPIRE NEVER ;
grant insert,delete,update,select on db144.* to 'test_r'@'%' ;

create user 'repl'@'%' IDENTIFIED with mysql_native_password by  'replpfhOTnWffQdQL3F3' ;
GRANT REPLICATION SLAVE ON *.* TO 'repl'@'%' ;
set sql_log_bin = 1;
复制代码
复制代码

十二、快捷方式设置

复制代码
复制代码
快捷方式
ln -s /usr/local/mysql/lib/libmysqlclient.so /usr/lib/
ln -s /usr/local/mysql/lib/libmysqlclient.so.21 /usr/lib/libmysqlclient.so.21
ln -s /usr/local/mysql/lib/libmysqlclient.so /usr/lib64/
ln -s /usr/local/mysql/lib/libmysqlclient.so.21 /usr/lib64/libmysqlclient.so.21


ln -s /data/mysql/mysql_3306/tmp/mysql_3306.sock /tmp/mysql.sock
ln -s /usr/local/mysql/bin/* /usr/bin/

cat >>~/.bashrc <<"EOF"
##########


alias mysql.3306.start="/usr/local/mysql/bin/mysqld_safe --defaults-file=/data/mysql/mysql_3306/my_3306.cnf &"
alias mysql.3306.stop="/usr/local/mysql/bin/mysqladmin -h127.0.0.1 -P 3306 -uroot -p'GJjumB6g4FcwdF3R6AZc' shutdown &"
alias mysql.3306.login="/usr/local/mysql/bin/mysql -h127.0.0.1 -P 3306 -uroot -p'GJjumB6g4FcwdF3R6AZc'"


##########
EOF


source  /root/.bash_profile

cat >>/etc/ld.so.conf <<"EOF"
/usr/local/mysql/lib
EOF

ldconfig 
mysql.3306.login
复制代码
复制代码

十三、MGR配置

13-1、第一台配置

复制代码

# MGR 第一台配置:
# 第一步:创建用于复制的用户

set sql_log_bin=0;
create user 'repuser'@'%' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'%';

create user 'repuser'@'127.0.0.1' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'127.0.0.1';

create user 'repuser'@'localhost' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'localhost';

set sql_log_bin=1;

# 第二步:配置复制所使用的用户

change master to master_user='repuser',master_password='JhXpMK44ju8Vp5bxvO2N' for channel 'group_replication_recovery';

# 第三步:安装mysql group replication这个插件


# 备注:如果在my.cnf里写写入 plugin_load="group_replication=group_replication.so" 这步就可以不用操作

install plugin group_replication soname 'group_replication.so';

# 通过show plugins;查看是否安装成功
show plugins;

# 第四步:建个群(官方点的说法就是初始化一个复制组

set global group_replication_bootstrap_group=on;
start group_replication;
set global group_replication_bootstrap_group=off;

select * from performance_schema.replication_group_members;


复制代码
# MGR  第一台配置:
# 第一步:创建用于复制的用户

set sql_log_bin=0;
create user 'repuser'@'%' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'%';

create user 'repuser'@'127.0.0.1' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'127.0.0.1';

create user 'repuser'@'localhost' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'localhost';

set sql_log_bin=1;

# 第二步:配置复制所使用的用户

change master to master_user='repuser',master_password='JhXpMK44ju8Vp5bxvO2N' for channel 'group_replication_recovery';

# 第三步:安装mysql group replication这个插件


# 备注:如果在my.cnf里写写入 plugin_load="group_replication=group_replication.so" 这步就可以不用操作

install plugin group_replication soname 'group_replication.so';

# 通过show plugins;查看是否安装成功
show plugins;

# 第四步:建个群(官方点的说法就是初始化一个复制组

set global group_replication_bootstrap_group=on;
start group_replication;
set global group_replication_bootstrap_group=off;

select * from performance_schema.replication_group_members;
复制代码
复制代码

13-2、第二台、第三台配置

复制代码
复制代码
##########################################################################
#MGR 配置其他从节点
#在所有从主机上的mysql中执行
# 第一步:创建用于复制的用户

set sql_log_bin=0;
create user 'repuser'@'%' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'%';

create user 'repuser'@'127.0.0.1' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'127.0.0.1';

create user 'repuser'@'localhost' identified by 'JhXpMK44ju8Vp5bxvO2N';

grant replication slave,replication client on *.* to 'repuser'@'localhost';

set sql_log_bin=1;

# 第二步:配置复制所使用的用户

change master to master_user='repuser',master_password='JhXpMK44ju8Vp5bxvO2N' for channel 'group_replication_recovery';

# 第三步:安装mysql group replication这个插件


# 备注:如果在my.cnf里写写入 plugin_load="group_replication=group_replication.so" 这步就可以不用操作

install plugin group_replication soname 'group_replication.so';

# 通过show plugins;查看是否安装成功
show plugins;

# 第四步:加入前面创建好的复制组
reset master;
start group_replication;
select * from performance_schema.replication_group_members;
#########################################################################################

# 检查状态
mysql> select * from performance_schema.replication_group_members;
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST | MEMBER_PORT | MEMBER_STATE | MEMBER_ROLE | MEMBER_VERSION |
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
| group_replication_applier | d955da6d-0048-11ea-b7b4-525400f4342d | bj-db-m1    |        3306 | ONLINE       | PRIMARY     | 8.0.18         |
| group_replication_applier | e050c34f-0048-11ea-917d-52540021fab9 | bj-db-m3    |        3306 | ONLINE       | SECONDARY   | 8.0.18         |
| group_replication_applier | e6c56347-0048-11ea-9e8b-5254007c241f | bj-db-m2    |        3306 | ONLINE       | SECONDARY   | 8.0.18         |
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
3 rows in set (0.00 sec)
复制代码
 正常情况下节点掉了可直接通过下面的方式拉起来
 
 
SET GLOBAL group_replication_bootstrap_group=ON;
start group_replication;
SET GLOBAL group_replication_bootstrap_group=off;
SELECT * FROM performance_schema.replication_group_members;
select @@read_only,@@super_read_only;
 
如果节点状态为下面的状态:
mysql> select * from performance_schema.replication_group_members;
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST | MEMBER_PORT | MEMBER_STATE | MEMBER_ROLE | MEMBER_VERSION |
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
| group_replication_applier | d955da6d-0048-11ea-b7b4-525400f4342d | bj-db-m1    |        3306 | ONLINE       | PRIMARY     | 8.0.18         |
| group_replication_applier | e050c34f-0048-11ea-917d-52540021fab9 | bj-db-m3    |        3306 | ONLINE       | UNREACHABLE   | 8.0.18         |
| group_replication_applier | e6c56347-0048-11ea-9e8b-5254007c241f | bj-db-m2    |        3306 | ONLINE       |UNREACHABLE  | 8.0.18 | 
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
3 rows in set (0.00 sec)
 
这个时候是无法直接拉起来的
需要先清除故障组组的配置:
set global group_replication_force_members='10.25.133.67:24901';
字保留正常组的配置
select * from performance_schema.replication_group_members;
 
然后直接在其他两个节点拉起来即可配置组:
或者直接启动加到组中
start group_replication;
复制代码

十四、单主切换到多主

复制代码
复制代码
######################################################################
# 单主切换到多主
# MGR切换模式需要重新启动组复制,因此需要在所有节点上先关闭组复制,
#设置 group_replication_single_primary_mode=OFF 等参数,再启动组复制。
1) 停止组复制(在所有MGR节点上执行):
stop group_replication; 
set global group_replication_single_primary_mode=OFF;
set global group_replication_enforce_update_everywhere_checks=ON;

2) 随便选择某个MGR节点执行 (比如这里选择在MGR-node1节点):
set global group_replication_recovery_get_public_key=1;
SET GLOBAL group_replication_bootstrap_group=ON;
START GROUP_REPLICATION;
SET GLOBAL group_replication_bootstrap_group=OFF;

3) 然后在其他的MGR节点执行 (这里指MGR-node2和MGR-node3节点上执行):
set global group_replication_recovery_get_public_key=1;
START GROUP_REPLICATION;

4) 查看MGR组信息 (在任意一个MGR节点上都可以查看)
SELECT * FROM performance_schema.replication_group_members;


# 可以看到所有MGR节点状态都是online,角色都是PRIMARY,MGR多主模式搭建成功。

##########################################################################
复制代码
复制代码

十五、多主切换回单主

复制代码
##########################################################################
# 多主切回单主模式
1) 停止组复制(在所有MGR节点上执行):
stop group_replication;
set global group_replication_enforce_update_everywhere_checks=OFF;
set global group_replication_single_primary_mode=ON;

2) 选择一个节点作为主节点, 在主节点上执行 (这里选择MGR-node1节点作为主节点)
SET GLOBAL group_replication_bootstrap_group=ON;
START GROUP_REPLICATION;
SET GLOBAL group_replication_bootstrap_group=OFF;
 
3) 在其他剩余的节点, 也就是从库节点上执行 (这里从库节点指的就是MGR-node2和MGR-node3):
START GROUP_REPLICATION;

4) 查看MGR组信息 (在任意一个MGR节点上都可以查看)
SELECT * FROM performance_schema.replication_group_members;
##########################################################################
复制代码
复制代码

十六、故障注意事项

复制代码
复制代码
# 故障注意点:
# 单主模式,恢复MGR-node1节点, 恢复后, 需要手动激活下该节点的组复制功能
# 如果节点发生故障, 在恢复后需要重新加入到MGR集群里, 正确的做法是:
STOP GROUP_REPLICATION;
START GROUP_REPLICATION;

# 如果某个节点挂了, 则其他的节点继续进行同步.
# 当故障节点恢复后, 只需要手动激活下该节点的组复制功能("START GROUP_REPLICATION;"),
# 即可正常加入到MGR组复制集群内并自动同步其他节点数据.


# 如果是i/o复制出现异常
# 确定数据无误后
# 查找主库的gtid情况
mysql> show global variables like '%gtid%' ;
+----------------------------------------------+-------------------------------------------------------+
| Variable_name                                | Value                                                 |
+----------------------------------------------+-------------------------------------------------------+
| binlog_gtid_simple_recovery                  | ON                                                    |
| enforce_gtid_consistency                     | ON                                                    |
| group_replication_gtid_assignment_block_size | 1000000                                               |
| gtid_executed                                | 58f6e65e-9309-11e9-9d88-525400184a0a:1-946050:1000003 |
| gtid_executed_compression_period             | 1000                                                  |
| gtid_mode                                    | ON                                                    |
| gtid_owned                                   |                                                       |
| gtid_purged                                  |                                                       |
| session_track_gtids                          | OFF                                                   |
+----------------------------------------------+-------------------------------------------------------+
9 rows in set (0.00 sec)

# 在有故障的从库中操作
stop GROUP_REPLICATION;
reset master;
set global gtid_purged='58f6e65e-9309-11e9-9d88-525400184a0a:1-946055:1000003';
START GROUP_REPLICATION;

# 添加白名单网段
stop group_replication;
set global group_replication_ip_whitelist="127.0.0.1/32,172.16.60.0/24,172.16.50.0/24,172.16.51.0/24";
start group_replication;
show variables like "group_replication_ip_whitelist";

# 一定要注意: 配置白名单前面一定要先关闭 Group Replication, 及先要执行"stop group_replication;"
复制代码

十七、MySQL Group Replication 节点状态详解

通过查询performance_schema下的replication_group_members表可以知道MGR集群中节点的状态:

mysql> desc performance_schema.replication_group_members;
+--------------+----------+------+-----+---------+-------+
| Field        | Type     | Null | Key | Default | Extra |
+--------------+----------+------+-----+---------+-------+
| CHANNEL_NAME | char(64) | NO   |     | NULL    |       |
| MEMBER_ID    | char(36) | NO   |     | NULL    |       |
| MEMBER_HOST  | char(60) | NO   |     | NULL    |       |
| MEMBER_PORT  | int(11)  | YES  |     | NULL    |       |
| MEMBER_STATE | char(64) | NO   |     | NULL    |       |
+--------------+----------+------+-----+---------+-------+
5 rows in set (0.01 sec)
  • CHANNEL_NAME : 显示的值永远为group_replication_applier
  • MEMBER_ID : 节点serer_uuid
  • MEMBER_PORT : 节点服务端口,取值为server_port指定的端口
  • MEMBER_HOST : 如果没有配置report_host选项,那么取值为机器的hostname,可以通过report_host配置指定具体的IP
  • MEMBER_STATE : 节点状态,取值下一节讨论

MEMBER_STATE取值

MEMBER_STATE字段显示当前节点的状态,根据官方文档,取值和介绍如下所示:

取值解释状态是否在集群内同步
ONLINE 表示该节点可正常提供服务 YES
RECOVERING 表示当前节点正在从其他节点恢复数据 YES
OFFLINE 表示GR插件已经加载,但是该节点不属于任何一个GR组 NO
ERROR 表示节点在recovery阶段出现错误或者从其他节点同步状态中出现错误 NO
UNREACHABLE 节点处于不可达状态,无法与之发生网络通讯 NO

从上表可以知道,只有ONLINE和RECOVERING两种状态会在集群中得到同步。这个状态同步是指状态在所有节点上面查询均能保持一致的意思。至于OFFLINE,ERROR和UNREABLE,做以下说明:

  • 只有在当前OFFLINE节点查询replication_group_members表才能得到OFFLINE状态,在其他节点上查询replication_group_members表,则一般没有该节点的状态(很好理解,因为OFFLINE节点已经不属于这个GR组了)
  • 只有在当前ERROR节点查询replication_group_members表才能得到ERROR状态,同上面的OFFLINE,在其他节点上查询也看不到该节点
  • 假设节点A与B网络通讯失败,那么在节点A上查询replication_group_members表,有可能得到B的状态为UNREACHABLE

那么,从状态是否自身可见或者其他节点可见的角度来区分,有如下区分,

状态对自身可见的有:

  • ONLINE
  • OFFLINE
  • RECOVERING
  • ERROR

状态在其他节点上可见的有:

  • ONLINE
  • RECOVERING
  • UNREACHABLE

节点状态转移

Group Replication Member State

当一个节点加进一个GR组,其状态首先变成RECOVERING,表示当前节点正处于集群恢复阶段,这个阶段下,节点会选择集群中一个节点(donor节点),利用传统的异步复制,做recovery。当数据能够成功追平,节点的状态将会变成ONLINE,这个过程中通过其他节点也可以看到该节点的状态,不管是RECOVERING还是最后的ONLINE。

假如该节点在RECOVERING阶段出现了异常(选donor进行复制失败 or 在donor追数据的过程中失败),那么该节点的状态将会变成ERROR,注意,这时候在其他节点上查询时,发现该RECOVERING节点已经从组里面被踢出。

另外,如果一个ONLINE节点失去与其他节点的通讯(可能因为节点crash或者网络异常),则该节点在其他节点上面查询到的状态将会是UNREACHABLE。如果这个UNREACHABLE节点在规定的超时时间内没有恢复过来,那么节点将会被踢出去。这个规定的超时时间是多少呢?下面会讲这个时间在集群失去这个节点是否可用的条件下是不一样的。

可疑的UNREACHABLE状态

前面提到,UNREACHABLE节点在规定的超时时间内如果没有恢复过来,那么节点将会被踢出去。这个规定的超时时间,取决于你集群失去这个节点下还是不是达到可用状态(之前的文章里面强调的2N + 1)。假设失去这个节点,集群仍然可用,那么这个UNREACHABLE的超时时间很短,几乎看不到这个状态;但是,如果失去这个节点后集群马上不可用,那么这个UNREACHABLE节点的超时时间,近似于无线大,将会一直处于UNREACHABLE!

以一个例子来验证:

3节点组成的集群,最开始3个节点均为ONLINE状态:

mysql> select * from performance_schema.replication_group_members;
+---------------------------+--------------------------------------+---------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST   | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+---------------+-------------+--------------+
| group_replication_applier | 748703ac-dbcc-11e6-a668-90e2bac5d924 | 10.202.44.215 |       24801 | ONLINE       |
| group_replication_applier | 8f8bc352-2566-11e7-aa5e-d4ae52ab91b3 | 10.202.44.214 |       24801 | ONLINE       |
| group_replication_applier | 9c09340a-e04c-11e6-9916-0024e869a606 | 10.202.44.213 |       24801 | ONLINE       |
+---------------------------+--------------------------------------+---------------+-------------+--------------+
3 rows in set (0.00 sec)

这时候,kill(注意是kill实例而不是正常down实例)掉其中的一个实例(10.202.44.214),通过其他可用节点查询到,那一个kill掉的实例从集群中被踢走了:

mysql> select * from performance_schema.replication_group_members;
+---------------------------+--------------------------------------+---------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST   | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+---------------+-------------+--------------+
| group_replication_applier | 748703ac-dbcc-11e6-a668-90e2bac5d924 | 10.202.44.215 |       24801 | ONLINE       |
| group_replication_applier | 9c09340a-e04c-11e6-9916-0024e869a606 | 10.202.44.213 |       24801 | ONLINE       |
+---------------------------+--------------------------------------+---------------+-------------+--------------+
2 rows in set (0.00 sec)

接下来我们再kill掉一个实例(10.202.44.213)

mysql> select * from performance_schema.replication_group_members;
+---------------------------+--------------------------------------+---------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST   | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+---------------+-------------+--------------+
| group_replication_applier | 748703ac-dbcc-11e6-a668-90e2bac5d924 | 10.202.44.215 |       24801 | ONLINE       |
| group_replication_applier | 9c09340a-e04c-11e6-9916-0024e869a606 | 10.202.44.213 |       24801 | UNREACHABLE  |
+---------------------------+--------------------------------------+---------------+-------------+--------------+
2 rows in set (0.00 sec)

这个时候,UNREACHABLE状态将一直持续,而且此时,集群不满足2N + 1,集群已经不可用(即使有主节点,主节点也是不可写的)。恢复主节点可写有两种方式:

  • 把前面kill掉的一个节点拉起来,加入GR组里面
  • 暴力地使用group_replication_force_members参数,强制设置节点组成一个新的GR组(强制剔除UNRECHABLE节点)

 

posted @   数据库小白(专注)  阅读(814)  评论(0编辑  收藏  举报
编辑推荐:
· 没有源码,如何修改代码逻辑?
· 一个奇形怪状的面试题:Bean中的CHM要不要加volatile?
· [.NET]调用本地 Deepseek 模型
· 一个费力不讨好的项目,让我损失了近一半的绩效!
· .NET Core 托管堆内存泄露/CPU异常的常见思路
阅读排行:
· 微软正式发布.NET 10 Preview 1:开启下一代开发框架新篇章
· 没有源码,如何修改代码逻辑?
· PowerShell开发游戏 · 打蜜蜂
· 在鹅厂做java开发是什么体验
· WPF到Web的无缝过渡:英雄联盟客户端的OpenSilver迁移实战
点击右上角即可分享
微信分享提示