Linux – What is the difference between repl-buffer and backlog in redis?

What is the difference between repl-buffer and backlog in redis?… here is a solution to the problem.

What is the difference between repl-buffer and backlog in redis?

In redis.conf, client-output-buffer-limit compared to repl-backlog-size?

redis master allocates a copy buffer for each slave, and I can set client-output-buffer-limit to 500M. repl-buff annotates the master command and runid’offset.

I don’t know the backlog impact in the Redis master role. Backlog records global offset, per dependent run, per different task time cost?

When slave note sends a psync command with a runid\last offset to the master, the master gets the data from the backlog or rel-buffer?

Solution

The copy buffer saves new updates to the master until an RDB snapshot of the master is transferred to the slave and loaded into its memory. After that, the data in the copy buffer is transferred to the slave server. Client-output-buffer-limit is the maximum size we allow the copy buffer to grow.

The backlog stores the data after it is disconnected from the server. So when the slave reconnects, we don’t need to do full synchronization.

So the answer to your question is copy buffers.

Related Problems and Solutions