class TokenAware

Returns a new instance of TokenAware

Note

If replicas are not shuffled (shuffle = false), then it is possibile to create hotspots in a write-heavy scenario, where most of the write requests will be handled by the same node(s). The default behavior of shuffling replicas helps mitigate this by universally distributing write load between replicas. However, it under-utilizes read caching and forces multiple replicas to cache the same read statements.

Parameters:

Name	Type	Details
wrapped_policy	LoadBalancing::Policy	actual policy to filter
shuffle	Boolean	(defaults to: true) (true) whether or not to shuffle the replicas

Delegates to wrapped policy

See Also:

• LoadBalancing::Policy#distance

Delegates to wrapped policy

See Also:

• LoadBalancing::Policy#host_found

Delegates to wrapped policy

See Also:

• LoadBalancing::Policy#host_up

Delegates to wrapped policy

See Also:

• LoadBalancing::Policy#host_down

Delegates to wrapped policy

See Also:

• LoadBalancing::Policy#host_lost

Specifications:

TokenAware#setup sets up wrapped policy

expect(policy).to receive(:setup).once.with(cluster)
subject.setup(cluster)