Hey all,
I have a production 3 node cluster with 250mm keys, and have been trying to build a test cluster with about 1/3 of the total data (so I can use it for testing).
What I tried to do was grab a rdb file from one of the production nodes, and use it as the seed for one of the test-cluster nodes, thinking I could redistribute the keys after that.
What I ended up with was three nodes with even slots, but all the data still on a single node. calling redis-trib rebalance or redis-trib reshard don't really seem to be able to help me rebalance the keys.
What can I do better? Does anyone have an idea how to get this to work?