关于自动故障转移模式下备库离线对主库造成影响问题
假如在自动故障转移模式下,镜像集群中两个镜像都是实时同步的Failover角色,当备库因为主机故障或者网络隔离等因素导致不可用,那么主库就收不到备库的ACKs,这个时候我看官方文档的解释是,主库会在Qos时间超时后,撤销备库日志传输的Active状态,那么在这个Qos时间段里(默认是8秒),主库是不会提交任何新的日志,那么遇到那些需要写或者更新的应用,应用会不会卡住8秒不动? 这对一个医院系统来说也是致命的。
希望哪位大神能为我解答一下,谢谢
translate
If in the automatic failover mode, the two images in the mirror cluster are both real-time synchronous failover roles.
When the standby database is unavailable due to host failure or network isolation,
the master database will not receive the acks of the standby database. At this time,
I see the explanation in the official document that the master database will revoke the active state of the log transmission
of the standby database after the QoS time expires, In this QoS time period (8 seconds by default), the main database will not submit any new logs.
Will the applications that need to be written or updated get stuck for 8 seconds? This is also fatal to a hospital system.