继续浏览精彩内容
慕课网APP
程序员的梦工厂
打开
继续
感谢您的支持,我会继续努力的
赞赏金额会直接到老师账户
将二维码发送给自己后长按识别
微信支付
支付宝支付

万答#12,MGR整个集群挂掉后,如何才能自动选主,不用手动干预

GreatSQL社区
关注TA
已关注
手记 112
粉丝 6
获赞 66

MGR整个集群挂掉后,如能才能自动选主,不用手动干预

当集群中所有节点都宕机后,集群再次启动后,能否自动选主?

1、所有节点都宕机后,集群重启,能否自动选主和启动

这是个来自群友的问题。

首先,MySQL服务利用 systemd 即可实现故障后自启动,注意下面这个配置即可:

[root@GreatSQL ~]# cat /usr/lib/systemd/system/greatsql.service
...
Restart=on-failure

其次,mysqld进程启动后,想要实现MGR的自动选主及自启动也是可以的,利用MySQL Shell即可,例如:

[root@GreatSQL ~]# mysqlsh --uri greatsql@yejr-mgr3:3306
...
-- 不管干啥,都先看 help,这是玩转Linux的必备素养,啥事不清楚都先找男人(man)
-- 注意到有这样的一个方法 rebootClusterFromCompleteOutage(),看起来没跑了
 MySQL  yejr-mgr3:3306 ssl  JS > \help dba
      rebootClusterFromCompleteOutage([clusterName][, options])
            Brings a cluster back ONLINE when all members are OFFLINE.

-- 跑一个试试看
 MySQL  yejr-mgr3:3306 ssl  JS > dba.rebootClusterFromCompleteOutage()
Restoring the default cluster from complete outage...

The instance 'yejr-mgr4:3306' was part of the cluster configuration.
Would you like to rejoin it to the cluster? [y/N]: y

The instance 'yejr-mgr2:3306' was part of the cluster configuration.
Would you like to rejoin it to the cluster? [y/N]: y

Dba.rebootClusterFromCompleteOutage: The active session instance (yejr-mgr3:3306) isn't the most updated in comparison with the ONLINE instances of the Cluster's metadata. Please use the most up to date instance: 'yejr-mgr4:3306'. (RuntimeError)

可以看到错误信息提示我们当前节点上没有最新的数据,不能直接启动MGR,错误信息中还提供了该去哪个节点启动的建议,所以我们改成在 yejr-mgr4 节点上执行拉起MGR:

[root@GreatSQL ~]# mysqlsh --uri greatsql@yejr-mgr3:3306
...
 MySQL  yejr-mgr4:3306 ssl  JS > dba.rebootClusterFromCompleteOutage()
Restoring the default cluster from complete outage...

The instance 'yejr-mgr3:3306' was part of the cluster configuration.
Would you like to rejoin it to the cluster? [y/N]: y

The instance 'yejr-mgr2:3306' was part of the cluster configuration.
Would you like to rejoin it to the cluster? [y/N]: y

yejr-mgr4:3306 was restored.
Rejoining 'yejr-mgr3:3306' to the cluster.
Rejoining instance 'yejr-mgr3:3306' to cluster 'GreatSQLMGR'...
The instance 'yejr-mgr3:3306' was successfully rejoined to the cluster.

Rejoining 'yejr-mgr2:3306' to the cluster.
Rejoining instance 'yejr-mgr2:3306' to cluster 'GreatSQLMGR'...
The instance 'yejr-mgr2:3306' was successfully rejoined to the cluster.

The cluster was successfully rebooted.

可以看到,MGR集群已经被正常启动了。

上面是利用MySQL Shell启动一个发生过故障的MGR集群,如果是手动的话该怎么办呢?

首先,在各个节点执行下面的SQL,确认各节点当前的事务执行情况:

-- yejr-mgr2节点
root@GreatSQL [none]> select RECEIVED_TRANSACTION_SET from performance_schema.replication_connection_status where 
channel_name = 'group_replication_applier' union all 
select variable_value from performance_schema.global_variables where 
variable_name = 'gtid_executed'\G
*************************** 1. row ***************************
RECEIVED_TRANSACTION_SET:
*************************** 2. row ***************************
RECEIVED_TRANSACTION_SET: 1c293e90-3bdc-11ec-bca1-525400e2078a:1-4537605,
4b7b3b88-3b13-11ec-86e9-525400e2078a:1

-- yejr-mgr3节点
...
*************************** 1. row ***************************
RECEIVED_TRANSACTION_SET:
*************************** 2. row ***************************
RECEIVED_TRANSACTION_SET: 1c293e90-3bdc-11ec-bca1-525400e2078a:1-4542304,
4b7b3b88-3b13-11ec-86e9-525400e2078a:1

-- yejr-mgr4节点
...
*************************** 1. row ***************************
RECEIVED_TRANSACTION_SET:
*************************** 2. row ***************************
RECEIVED_TRANSACTION_SET: 1c293e90-3bdc-11ec-bca1-525400e2078a:1-4652391,
4b7b3b88-3b13-11ec-86e9-525400e2078a:1

从上面的结果可以看到,yejr-mgr4 节点上已执行完的事务GTID值最大:4652391 > 4542304 > 4537605,因此应该选择 yejr-mgr4 节点作为 Primary 节点。

将该节点设置为引导模式,然后启动MGR服务:

[root@GreatSQL ~]# mysql -hyejr-mgr4 -P3306 -ugreatsql -p
...
greatsql@mgr4:3306 [(none)]>set global group_replication_bootstrap_group=ON;

greatsql@mgr4:3306 [(none)]>start group_replication;

-- 启动完MGR后,记得立即将其设置为OFF
greatsql@mgr4:3306 [(none)]>set global group_replication_bootstrap_group=OFF;

在其他节点上,则直接启动MGR服务即可,切记无需再次设置引导模式,否则它就会变成一个全新的MGR集群的Primary节点了。

好了,自动、手动两种方式拉起一个故障MGR集群方法都介绍完毕了。

Enjoy GreatSQL :)

打开App,阅读手记
1人推荐
发表评论
随时随地看视频慕课网APP