2

我们在其中一个环境中有一个 MySQL innodb 集群。集群中的一个节点崩溃了。虽然，我们能够使崩溃的节点联机，但我们无法将其加入集群。

有人可以帮助恢复/恢复节点并将其加入集群。我们尝试使用“dba.rebootClusterFromCompleteOutage()”，但没有帮助。

配置：MySQL 5.7.24社区版，CentOS 7，标准三节点innodb集群

集群状态：

MySQL  NODE02:3306 ssl  JS > var c=dba.getCluster()
MySQL  NODE02:3306 ssl  JS > c.status()
{
    "clusterName": "QACluster",
    "defaultReplicaSet": {
        "name": "default",
        "primary": "NODE03:3306",
        "ssl": "REQUIRED",
        "status": "OK_NO_TOLERANCE",
        "statusText": "Cluster is NOT tolerant to any failures. 1 member is not active",
        "topology": {
            "NODE02:3306": {
                "address": "NODE02:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE03:3306": {
                "address": "NODE03:3306",
                "mode": "R/W",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE01:3306": {
                "address": "NODE01:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "(MISSING)"
            }
        }
    },
    "groupInformationSourceMember": "mysql://clusterAdmin@NODE03:3306"
}

mysql错误日志中记录的错误：

2019-03-04T23:49:36.970839Z 3624 [Note] Slave SQL thread for channel 'group_replication_recovery' initialized, starting replication in log 'FIRST' at position 0, relay log './NODE01-relay-bin-group_replication_recovery.000001' position: 4
2019-03-04T23:49:36.985336Z 3623 [Note] Slave I/O thread for channel 'group_replication_recovery': connected to master 'mysql_innodb_cluster_r0429584112@NODE02:3306',replication started in log 'FIRST' at position 4
2019-03-04T23:49:36.988164Z 3623 [ERROR] Error reading packet from server for channel 'group_replication_recovery': The slave is connecting using CHANGE MASTER TO MASTER_AUTO_POSITION = 1, but the master has purged binary logs containing GTIDs that the slave requires. (server_errno=1236)
2019-03-04T23:49:36.988213Z 3623 [ERROR] Slave I/O for channel 'group_replication_recovery': Got fatal error 1236 from master when reading data from binary log: 'The slave is connecting using CHANGE MASTER TO MASTER_AUTO_POSITION = 1, but the master has purged binary logs containing GTIDs that the slave requires.', Error_code: 1236
2019-03-04T23:49:36.988226Z 3623 [Note] Slave I/O thread exiting for channel 'group_replication_recovery', read up to log 'FIRST', position 4
2019-03-04T23:49:36.988286Z 41 [Note] Plugin group_replication reported: 'Terminating existing group replication donor connection and purging the corresponding logs.'
2019-03-04T23:49:36.988358Z 3624 [Note] Error reading relay log event for channel 'group_replication_recovery': slave SQL thread was killed
2019-03-04T23:49:36.988435Z 3624 [Note] Slave SQL thread for channel 'group_replication_recovery' exiting, replication stopped in log 'FIRST' at position 0
2019-03-04T23:49:37.016864Z 41 [Note] 'CHANGE MASTER TO FOR CHANNEL 'group_replication_recovery' executed'. Previous state master_host='NODE02', master_port= 3306, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='<NULL>', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''.
2019-03-04T23:49:37.030769Z 41 [ERROR] Plugin group_replication reported: 'Maximum number of retries when trying to connect to a donor reached. Aborting group replication recovery.'
2019-03-04T23:49:37.030798Z 41 [Note] Plugin group_replication reported: 'Terminating existing group replication donor connection and purging the corresponding logs.'
2019-03-04T23:49:37.051169Z 41 [Note] 'CHANGE MASTER TO FOR CHANNEL 'group_replication_recovery' executed'. Previous state master_host='<NULL>', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='<NULL>', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''.
2019-03-04T23:49:37.069184Z 41 [ERROR] Plugin group_replication reported: 'Fatal error during the Recovery process of Group Replication. The server will leave the group.'
2019-03-04T23:49:37.069304Z 41 [Note] Plugin group_replication reported: 'Going to wait for view modification'
2019-03-04T23:49:40.336938Z 0 [Note] Plugin group_replication reported: 'Group membership changed: This member has left the group.'
4

2 回答 2

5

我执行以下操作以从备份中恢复故障节点并能够恢复集群状态。

1）以下是其中一个节点发生故障（NODE01）时集群的状态。 

 MySQL  NODE02:3306 ssl  JS > var c=dba.getCluster()
 MySQL  NODE02:3306 ssl  JS > c.status()
{
    "clusterName": "QACluster",
    "defaultReplicaSet": {
        "name": "default",
        "primary": "NODE03:3306",
        "ssl": "REQUIRED",
        "status": "OK_NO_TOLERANCE",
        "statusText": "Cluster is NOT tolerant to any failures. 1 member is not active",
        "topology": {
            "NODE02:3306": {
                "address": "NODE02:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE03:3306": {
                "address": "NODE03:3306",
                "mode": "R/W",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE01:3306": {
                "address": "NODE01:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "(MISSING)"
            }
        }
    },
    "groupInformationSourceMember": "mysql://clusterAdmin@NODE03:3306"
}

2）使用以下命令从主节点（健康节点）获取mysqldump。

[root@NODE03 db_backup]# mysqldump --all-databases --add-drop-database --single-transaction --triggers --routines --port=mysql_port --user=root -p > /db_backup/mysql_dump_03062019.sql
Enter password:
Warning: A partial dump from a server that has GTIDs will by default include the GTIDs of all transactions, even those that changed suppressed parts of the database. If you don't want to restore GTIDs, pass --set-gtid-purged=OFF. To make a complete dump, pass --all-databases --triggers --routines --events.

3) 执行以下步骤以从集群中删除故障节点。

 MySQL  NODE03:3306 ssl  JS > var c=dba.getCluster()
 MySQL  NODE03:3306 ssl  JS > c.rescan()
Rescanning the cluster...

Result of the rescanning operation:
{
    "defaultReplicaSet": {
        "name": "default",
        "newlyDiscoveredInstances": [],
        "unavailableInstances": [
            {
                "host": "NODE01:3306",
                "label": "NODE01:3306",
                "member_id": "e2aa897d-1828-11e9-85b3-00505692188c"
            }
        ]
    }
}

The instance 'NODE01:3306' is no longer part of the HA setup. It is either offline or left the HA group.
You can try to add it to the cluster again with the cluster.rejoinInstance('NODE01:3306') command or you can remove it from the cluster configuration.
Would you like to remove it from the cluster metadata? [Y/n]: Y
Removing instance from the cluster metadata...

The instance 'NODE01:3306' was successfully removed from the cluster metadata.

 MySQL  NODE03:3306 ssl  JS > c.status()
{
    "clusterName": "QACluster",
    "defaultReplicaSet": {
        "name": "default",
        "primary": "NODE03:3306",
        "ssl": "REQUIRED",
        "status": "OK_NO_TOLERANCE",
        "statusText": "Cluster is NOT tolerant to any failures.",
        "topology": {
            "NODE02:3306": {
                "address": "NODE02:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE03:3306": {
                "address": "NODE03:3306",
                "mode": "R/W",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            }
        }
    },
    "groupInformationSourceMember": "mysql://clusterAdmin@NODE03:3306"
}

4) 如果组复制仍在故障节点上运行，则停止组复制。

mysql> STOP GROUP_REPLICATION;
Query OK, 0 rows affected (1.01 sec)

5) 在故障节点上重置“gtid_executed”。

mysql> show global variables like 'GTID_EXECUTED';
+---------------+--------------------------------------------------------------------------------------------+
| Variable_name | Value                                                                                      |
+---------------+--------------------------------------------------------------------------------------------+
| gtid_executed | 01f27b9c-182a-11e9-a199-00505692188c:1-14134172,
e2aa897d-1828-11e9-85b3-00505692188c:1-12 |
+---------------+--------------------------------------------------------------------------------------------+
1 row in set (0.01 sec)

mysql> reset master;
Query OK, 0 rows affected (0.02 sec)

mysql> reset slave;
Query OK, 0 rows affected (0.02 sec)

mysql> show global variables like 'GTID_EXECUTED';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| gtid_executed |       |
+---------------+-------+
1 row in set (0.00 sec)

6) 在故障节点上禁用“super_readonly_flag”。

mysql> SELECT @@global.read_only, @@global.super_read_only;
+--------------------+--------------------------+
| @@global.read_only | @@global.super_read_only |
+--------------------+--------------------------+
|                  1 |                        1 |
+--------------------+--------------------------+
1 row in set (0.00 sec)

mysql> SET GLOBAL super_read_only = 0;
Query OK, 0 rows affected (0.00 sec)

mysql> SELECT @@global.read_only, @@global.super_read_only;
+--------------------+--------------------------+
| @@global.read_only | @@global.super_read_only |
+--------------------+--------------------------+
|                  1 |                        0 |
+--------------------+--------------------------+
1 row in set (0.00 sec)

7) 将 mysqldump 从 master 恢复到故障节点。

[root@E2LXQA1ALFDB01 db_backup]# mysql -uroot -p < mysql_dump_03062019.sql

8) 一旦恢复完成，在故障节点上启用“super_readonly_flag”。

mysql> SELECT @@global.read_only, @@global.super_read_only;
+--------------------+--------------------------+
| @@global.read_only | @@global.super_read_only |
+--------------------+--------------------------+
|                  1 |                        0 |
+--------------------+--------------------------+
1 row in set (0.00 sec)

mysql> SET GLOBAL super_read_only = 1;
Query OK, 0 rows affected (0.00 sec)


mysql> SELECT @@global.read_only, @@global.super_read_only;
+--------------------+--------------------------+
| @@global.read_only | @@global.super_read_only |
+--------------------+--------------------------+
|                  1 |                        1 |
+--------------------+--------------------------+
1 row in set (0.00 sec)

9）最后将故障节点添加回innodb集群。

MySQL  NODE03:3306 ssl  JS > c.addInstance('clusterAdmin@NODE01:3306');
A new instance will be added to the InnoDB cluster. Depending on the amount of
data on the cluster this might take from a few seconds to several hours.

Adding instance to the cluster ...

Please provide the password for 'clusterAdmin@NODE01:3306': *******************
Save password for 'clusterAdmin@NODE01:3306'? [Y]es/[N]o/Ne[v]er (default No):
Validating instance at NODE01:3306...

This instance reports its own address as NODE01
WARNING: The following tables do not have a Primary Key or equivalent column:
ephesoft.dlf, report.correction_type, report.field_details_ag, report_archive.correction_type, report_archive.field_details_ag, report_archive.global_data_ag

Group Replication requires tables to use InnoDB and have a PRIMARY KEY or PRIMARY KEY Equivalent (non-null unique key). Tables that do not follow these requirements will be readable but not updateable when used with Group Replication. If your applications make updates (INSERT, UPDATE or DELETE) to these tables, ensure they use the InnoDB storage engine and have a PRIMARY KEY or PRIMARY KEY Equivalent.

Instance configuration is suitable.
WARNING: On instance 'NODE01:3306' membership change cannot be persisted since MySQL version 5.7.24 does not support the SET PERSIST command (MySQL version >= 8.0.11 required). Please use the .configureLocalInstance command locally to persist the changes.
WARNING: On instance 'NODE02:3306' membership change cannot be persisted since MySQL version 5.7.24 does not support the SET PERSIST command (MySQL version >= 8.0.11 required). Please use the .configureLocalInstance command locally to persist the changes.
WARNING: On instance 'NODE03:3306' membership change cannot be persisted since MySQL version 5.7.24 does not support the SET PERSIST command (MySQL version >= 8.0.11 required). Please use the .configureLocalInstance command locally to persist the changes.
The instance 'clusterAdmin@NODE01:3306' was successfully added to the cluster.


 MySQL  NODE03:3306 ssl  JS > c.status()
{
    "clusterName": "QACluster",
    "defaultReplicaSet": {
        "name": "default",
        "primary": "NODE03:3306",
        "ssl": "REQUIRED",
        "status": "OK",
        "statusText": "Cluster is ONLINE and can tolerate up to ONE failure.",
        "topology": {
            "NODE01:3306": {
                "address": "NODE01:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE02:3306": {
                "address": "NODE02:3306",
                "mode": "R/O",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            },
            "NODE03:3306": {
                "address": "NODE03:3306",
                "mode": "R/W",
                "readReplicas": {},
                "role": "HA",
                "status": "ONLINE"
            }
        }
    },
    "groupInformationSourceMember": "mysql://clusterAdmin@NODE03:3306"
}
于 2019-03-09T04:31:07.193 回答
2

2019-03-04T23:49:36.988213Z 3623 [错误] 通道“group_replication_recovery”的从站 I/O：从主站读取二进制日志数据时出现致命错误 1236：“从站正在使用 CHANGE MASTER TO MASTER_AUTO_POSITION = 1 进行连接，但主服务器已清除包含从服务器所需的 GTID 的二进制日志。'，错误代码：1236

让我们简单来说：这意味着您有 2 台服务器（S1，S2），一台崩溃（S2）。当 S2 崩溃时，您在 S1 上执行了一些操作（事务 T1，T2），但在某个时间点，S1 上的二进制日志被清除，它们包含 T1 的信息。

最重要的是，S2 无法加入该组，因为它知道它缺少 T1，但是没有包含该信息的 binlog 文件可用于在恢复期间传输该信息。

抱歉，这里没有魔法命令，那只是运气不好。目前，您最好的选择是使用诸如转储之类的工具或该组中的其他工具为 S2 提供来自 S1 的数据。

将来，您应该重新检查您的 binlog 清除策略，并尽可能加快崩溃机器的恢复速度。

于 2019-03-07T11:22:39.133 回答