question

When a node is added to the cluster, it will receive new mutations (write requests) for the token range it will eventually own while it is bootstrapping. This means that the existing owner of the token range and the new node is receiving new write requests at the same time during the bootstrap.

But the new node will not accept read requests until the B has completed. Cheers!

一个新节点在加入到cluster的时候，它会被标记成UJ状态（意味着UP and Joining）。处于Joining状态的节点已经被安排好了它将要负责的Token Range，所以任何涉及到这个新Token Range的写请求都会被多写一份到这个新节点，在Joining完成之前(也就是Bootstrapping期间)，已有的负责这些Token Range的节点仍然会收到所有的写请求。这意味着token range的现有所有者和新节点在bootstrap期间会同时接收新的写请求。

但是新节点将在bootstrap完成之前不接受读取请求。