Implementation of vnode will provide us with lot of advantages in bootstrapping and operating a big cluster.
Some Problems which VNodes can solve:
- When we have a balanced large cluster doubling a live cluster is not an very good option and often this makes us to over provision.
- The above is true when we want to shrink a cluster too.
- Currently in my organization often we have use cases where we have to refresh production cluster's into a dev clusters. There is a strong need where we can have different sized clusters and data' can be fork lift into and out of it without over engineering or complicating the recovery process.
- In some cases bootstrapping a node which has a large amount of data might be faster without over loading the neighbors.
I am not sure if i know all the changes which needs to be performed to get to that state, but this ticket will get us started.