我的DC有4个节点，查看节点status都是正常的，每小时会有1000多同时数据写入，经常性的出现 hinted handoff，节点之间转移数据，导致客户端请求写入Request timed out，写入策略是one，请问是什么原因呢？
Hinted handoff不是写超时的原因。 这只是集群中节点无响应的征兆，因此hints最可能由coordinator存储，因为节点无法跟上write requests的速度。
The hinted handoff is not the cause of the write timeouts. It is just a symptom of nodes in your cluster becoming unresponsive so hints are stored by the coordinator most likely because the nodes cannot keep up with the write requests.
It also does not matter that writes have a consistency level of ONE. All mutations are sent by the coordinator to all replicas. Cheers!
5 People are following this question.