最常见的note日志以下三种

第一种，Aborted connection . ：

[Note] Aborted connection 61 to db: ‘unconnected’ user: ‘mha’ host: ‘197.xx.xx.xx(Got an error reading communication packets)

原因是因为网络等原因导致。

[Note] Aborted connection 61 to db: ‘unconnected’ user: ‘mha’ host: ‘197.xx.xx.xx(Got timeout reading communication packets)

原因是会话的idle时间达到了数据库指定的timeout时间。

第二种：SLAVE多线程同步的信息

2016-03-14T15:48:26.432150Z 73 [Note]Multi-threaded slave statistics for channel ”: seconds elapsed = 121; eventsassigned = 100374529; worker queues filled over

overrun level = 0; waited due aWorker queue full = 0; waited due the total size = 0; waited at clock conflicts= 1451875661700 waited (count) when Workers occupied = 3211993 waited whenWorkers occupied = 445032386000

events assigned：总共有多少个event被分配执行，计的是总数。

worker queues filled over overrun level：多线程同步中，worker 的私有队列长度超长的次数，计的是总数。

waited due a Worker queue full ：因为worker的队列超长而产生等待的次数，计的是总数。

waited due the total size ：超过最大size的次数，这个由参数slave_pending_jobs_size_max 指定。

waited at clock conflicts ：因为逻辑时间产生冲突的等待时间，单位是纳秒。

waited (count) when Workers occupied ：因为workder被占用而出现等待的次数。（总计值）。

waited when Workersoccupied ：因为workder被占用而出现等待的总时间，总计值，单位是纳秒。

第三种：page_cleaner线程的输出日志

[Note] InnoDB:page_cleaner: 1000ms intended loop took 4750ms.The

settings might not be optimal. (flushed=1519 and evicted=0, during the time.)

本次刷新时间 - 上次刷新时间 > 1秒(睡眠时间)+3秒 则报警告,一般是IO能力不足

参数优化(目的在于减少每次刷新的量，让每次刷新块数更加平均)：

innodb_io_capacity 应该降低

innodb_max_dirty_pages_pct 应该降低

innodb_max_dirty_pages_pct_lwm 如果设置了应该考虑降低

innodb_io_capacity_max 降低