Active Failover redirects should not follow Location header

In active failover all requests must be made against the leader. Requests to followers will be rejected with HTTP 503. The response will contain a Location header that indicates where the actual leader is. The Java driver automatically retries sending the request to the endpoint in the Location header.

Due to server limitations, Active Failover deployment does not allow distinguishing internal and external names for db endpoints.
Therefore the Location header returned from the followers replies refers to the leader's internal name/IP address, which could not be reachable from the driver.

To work around it, the driver requests should not be redirected to the endpoint specified in the Location header, but retried with other hosts from the driver configuration hostlist, which could be manually configured with external names/IPs.

In case the internal names/IPs are reachable from the driver, the user could configure the driver with acquireHostList to discover all the db endpoints.

No files, tests, or entry points are named in the issue. Locate the Java driver's active-failover redirect and retry handling, then verify that failed requests retry configured hosts instead of following the response Location header, while preserving acquireHostList behavior.