參考官方文檔，分析了一下https://dev.mysql.com/doc/refman/5.7/en/communication-errors.html
發現這個描述跟遇到的情況比較類似，即因爲客戶端空閒時間超過了wait_timeout或interactive_timeout超時導致。
If a client successfully connects but later disconnects improperly or is terminated,
the server increments the Aborted_clients status variable, and logs an Aborted connection message to the error log.
The cause can be any of the following:
The client program did not call mysql_close() before exiting.
The client had been sleeping more than wait_timeout or interactive_timeout seconds without issuing any requests to the server. See Section 5.1.5, “Server System Variables”.
The client program ended abruptly in the middle of a data transfer.
mysql> select @@version;
+------------+
| @@version |
+------------+
| 5.7.21-log |
+------------+
1 row in set (0.00 sec)
錯誤日誌
tail -f error.log
[Note] Aborted connection 10685 to db: 'zabbix' user: 'monitor' host: '192.168.xxx.xxx' (Got timeout reading communication packets)
當日志中出現錯誤時，Aborted_clients數量也在增加
mysql> show status like 'Aborted%';
+------------------+-------+
| Variable_name | Value |
+------------------+-------+
| Aborted_clients | 1964 |
| Aborted_connects | 69 |
+------------------+-------+
2 rows in set (0.01 sec)
通過如下SQL預測下次出現Aborted connection的時間
select sysdate(), id,SUBSTRING_INDEX(host,':',1) host,time, DATE_ADD(sysdate() ,INTERVAL @@wait_timeout-time second) 'Aborted Time'
from information_schema.PROCESSLIST where id>100 and time>1000 order by time desc ,id;
+---------------------+-------+----------------+------+---------------------+
| sysdate() | id | host | time | Aborted Time |
+---------------------+-------+----------------+------+---------------------+
| 2018-03-30 16:10:52 | 10786 | 192.168.x.191 | 1800 | 2018-03-30 16:10:52 |
| 2018-03-30 16:10:52 | 10787 | 192.168.x.117 | 1800 | 2018-03-30 16:10:52 |
| 2018-03-30 16:10:52 | 10748 | 192.168.x.221 | 1735 | 2018-03-30 16:11:57 |
| 2018-03-30 16:10:52 | 10692 | 192.168.x.221 | 1686 | 2018-03-30 16:12:46 |
| 2018-03-30 16:10:52 | 10777 | 192.168.x.221 | 1102 | 2018-03-30 16:22:30 |
+---------------------+-------+----------------+------+---------------------+
5 rows in set (0.01 sec)
檢查error.log發現日誌出現的時間與SQL中得到的Aborted Time吻合.
2018-03-30T16:10:52.661416+08:00 10786 [Note] Aborted connection 10786 to db: 'db01' user: 'apps' host: '192.168.0.191' (Got timeout reading communication packets)
2018-03-30T16:10:52.668516+08:00 10787 [Note] Aborted connection 10787 to db: 'db02' user: 'apps' host: '192.168.0.117' (Got timeout reading communication packets)
2018-03-30T16:11:57.564463+08:00 10748 [Note] Aborted connection 10748 to db: 'db03' user: 'apps' host: '192.168.x.221' (Got timeout reading communication packets)
解決辦法：
mysql> set global log_error_verbosity=2;
Query OK, 0 rows affected (0.00 sec)
參考文檔：
https://dev.mysql.com/doc/refman/5.7/en/server-system-variables.html#sysvar_log_error_verbosity