Question Multi-Master/Active Replica (Issue rbd related?)

we were trying to use multi master cluster or active replica but our db is quite huge, few hundred million keys. We noticed that, despite rbd bg save disabled, rbds are created and send over to the other nodes, is this how it works in keydb? It seems weird. The Problem is that sending and important that rbd is slower than the time it takes to start another rbd process …so it never works. This seems not normal?
Shouldnt data be send directly between nodes without any rbd dumbs?
And yes rbd bg save is disabled.

Thanks for any hints

The way Multi-Master/Active Replication works is that whenever KeyDB servers syncs with another, a RDB background save on each server is performed first. After this save, the KeyDB servers load the data from the RDB save to transfer it to another KeyDB server.

You may want to use a ring topology rather than a mesh topology. Also try to add a KeyDB servers one at a time. Wait for each new KeyDB server to finish sync’ing up with the rest of the Servers before adding another one.