Geek Answers Handbook

Why is cluster redirection not automatic?

When I add node to the redis cluster, it has 0 hash slots. Why doesn't the redis cluster automatically perform a resharding operation to make the new node fully functional?

+5
source share
3 answers

As you can see here , redis now supports automatic partitioning.

+1
source

The process of adding a node consists of two steps:

  • Introduce node to other nodes through CLUSTER MEET so that all nodes start communicating via the cluster bus
  • Make node act as Master through CLUSTER ADDSLOTS or as subordinate through CLUSTER REPLICATE

Separation helps simplify the execution of commands.

Automatic forwarding is part of the Redis 4.2 roadmap

0
source

Given my experience, I do not want automatic reshard.

I was dealing with the fact that some node has a high read throughput (100 kbps), so I add new nodes to protect only these nodes with high load in order to reduce pressure.

You may ask why the load is different? Because we use a hash tag (for example, {user} 123456) to provide the same type of data stored on the same node.

so automatic rehard is useless.

0
source

More articles:


All Articles