Problem
HostConnection.return_connection() still runs down-handling for defunct or closed connections after that pool has stopped being the current pool for the host. If a topology refresh replaced the pool or changed the host endpoint first, the retired pool can still call host.signal_connection_failure() and drive cluster down handling against the current host state.
Impact
A failure from a stale pool can incorrectly mark the replacement endpoint down, tear down the healthy pool, or start unnecessary reconnection work.
Problem
HostConnection.return_connection()still runs down-handling for defunct or closed connections after that pool has stopped being the current pool for the host. If a topology refresh replaced the pool or changed the host endpoint first, the retired pool can still callhost.signal_connection_failure()and drive cluster down handling against the current host state.Impact
A failure from a stale pool can incorrectly mark the replacement endpoint down, tear down the healthy pool, or start unnecessary reconnection work.