Why three nodes in the cluster have different ranges?

I have started a cluster with three nodes.

I’m confused that why these nodes have diffrent ranges, why sysbench performance test result is much better when connecting to 101 than connecting to 102.

The owner (lease holder) of each range moves to follow where most of the load is. It might take a little while for the lease holders to move to another node. How long are you running the sysbench test for?

Did you restart any of the nodes? This looks like it may be a reporting bug (it’s similar to https://github.com/cockroachdb/cockroach/issues/19294). If you could run cockroach debug zip --insecure and share the resulting file with us, it should help us figure out what the source of the problem is.