1

我在 3 节点 galera-cluster 前面有 2 个 proxysql(2.0.13) + keepalived，并且 2 个 galera-nodes 在我的主机组 10 中显示为“SHUNNED”。我不知道为什么它们显示为“SHUNNED”他们永远不会恢复。

所以我希望这里有人可以帮助我将回避的服务器带回循环中。

我试过的： - 设置 var mysql-shun_on_failures = 5000 & mysql-shun_recovery_time_sec = 10 - 重启 proxysql 节点

我已经连接到主机组 10，并进行了一些查询，以触发恢复，但它们仍然被回避。

这是proxysql的一些信息

Servers:
Admin> select * from runtime_mysql_servers;
+--------------+-----------+------+-----------+---------+--------+-------------+-----------------+---------------------+---------+----------------+---------+
| hostgroup_id | hostname  | port | gtid_port | status  | weight | compression | max_connections | max_replication_lag | use_ssl | max_latency_ms | comment |
+--------------+-----------+------+-----------+---------+--------+-------------+-----------------+---------------------+---------+----------------+---------+
| 10           | 10.4.4.21 | 3306 | 0         | SHUNNED | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 20           | 10.4.4.21 | 3306 | 0         | ONLINE  | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 20           | 10.4.4.22 | 3306 | 0         | ONLINE  | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 30           | 10.4.4.23 | 3306 | 0         | ONLINE  | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 30           | 10.4.4.22 | 3306 | 0         | ONLINE  | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 30           | 10.4.4.21 | 3306 | 0         | ONLINE  | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 10           | 10.4.4.23 | 3306 | 0         | ONLINE  | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
| 10           | 10.4.4.22 | 3306 | 0         | SHUNNED | 1      | 0           | 1000            | 0                   | 0       | 0              |         |
+--------------+-----------+------+-----------+---------+--------+-------------+-----------------+---------------------+---------+----------------+---------+


Hostgroups:
Admin> select * from runtime_mysql_galera_hostgroups;
+------------------+-------------------------+------------------+-------------------+--------+-------------+-----------------------+-------------------------+--------------------+
| writer_hostgroup | backup_writer_hostgroup | reader_hostgroup | offline_hostgroup | active | max_writers | writer_is_also_reader | max_transactions_behind | comment            |
+------------------+-------------------------+------------------+-------------------+--------+-------------+-----------------------+-------------------------+--------------------+
| 10               | 20                      | 30               | 9999              | 1      | 1           | 1                     | 30                      | galera-hostgroup-1 |
+------------------+-------------------------+------------------+-------------------+--------+-------------+-----------------------+-------------------------+--------------------+

Status:
Admin> select hostgroup,srv_host,status,ConnUsed,MaxConnUsed,Queries,Latency_us from stats.stats_mysql_connection_pool order by srv_host;
+-----------+-----------+---------+----------+-------------+---------+------------+
| hostgroup | srv_host  | status  | ConnUsed | MaxConnUsed | Queries | Latency_us |
+-----------+-----------+---------+----------+-------------+---------+------------+
| 10        | 10.4.4.21 | SHUNNED | 0        | 0           | 0       | 289        |
| 30        | 10.4.4.21 | ONLINE  | 0        | 0           | 0       | 289        |
| 20        | 10.4.4.21 | ONLINE  | 0        | 0           | 0       | 289        |
| 10        | 10.4.4.22 | SHUNNED | 0        | 0           | 0       | 281        |
| 30        | 10.4.4.22 | ONLINE  | 0        | 0           | 0       | 281        |
| 20        | 10.4.4.22 | ONLINE  | 0        | 0           | 0       | 281        |
| 10        | 10.4.4.23 | ONLINE  | 1        | 1           | 106     | 277        |
| 30        | 10.4.4.23 | ONLINE  | 0        | 0           | 0       | 277        |
+-----------+-----------+---------+----------+-------------+---------+------------+

My ping-log:
Admin> select * from monitor.mysql_server_ping_log order by time_start_us DESC limit 0,10;
+-----------+------+------------------+----------------------+------------+
| hostname  | port | time_start_us    | ping_success_time_us | ping_error |
+-----------+------+------------------+----------------------+------------+
| 10.4.4.23 | 3306 | 1597916943398296 | 257                  | NULL       |
| 10.4.4.21 | 3306 | 1597916943298413 | 232                  | NULL       |
| 10.4.4.22 | 3306 | 1597916943198499 | 276                  | NULL       |
| 10.4.4.22 | 3306 | 1597916935355114 | 220                  | NULL       |
| 10.4.4.23 | 3306 | 1597916935277079 | 259                  | NULL       |
| 10.4.4.21 | 3306 | 1597916935199034 | 232                  | NULL       |
| 10.4.4.22 | 3306 | 1597916927359456 | 175                  | NULL       |
| 10.4.4.21 | 3306 | 1597916927278902 | 242                  | NULL       |
| 10.4.4.23 | 3306 | 1597916927198335 | 302                  | NULL       |
| 10.4.4.21 | 3306 | 1597916919324515 | 189                  | NULL       |
+-----------+------+------------------+----------------------+------------+

谢谢你们..

4

1 回答 1

0

我找到了他们被标记为“回避”的原因：

我的主机组配置的 max_writers = 1 - 这意味着当时只有我的 3 个作家（在我的作家主机组中）应该可用，另外 2 个应该待命，所以如果当前作家可以接管应该失败。

这里解释得很好——在“ProxySQL 2.x Support for Galera Cluster”下查看 https://severalnines.com/database-blog/how-run-and-configure-proxysql-20-mysql-galera-cluster-docker .

于 2020-08-24T14:24:56.437 回答