The netty_client_request_timeout in the dse.yaml file helps to prevent timeouts against queries that may run for too long. In DCs where RF= number of nodes (RF=ALL), this setting does not have any effect
When a long running query is run against the DC which has RF=ALL, the query will just run to completion and never hit this timeout threshold.
This is expected behaviour; the netty_client_request_timeout is only applicable to networked connections. In the case of a DC where RF=ALL then all nodes hold all the data, and therefore this setting is not effective.
If it is important to have the timeout in use then increasing the node count or reducing the RF of the user keyspace will be necessary.