Theses changes allows the retry mechanism to retry the request if any error other than timeout occurs.
In case of a cluster, this will ensure the connection where the error occured is marked as dead and request will be retried on a new host provided by the pool.
Theses changes allows the retry mechanism to retry the request if any error other than timeout occurs. In case of a cluster, this will ensure the connection where the error occured is marked as dead and request will be retried on a new host provided by the pool.
Relates to #2