Yes, instead of relying on specific exceptions to determine when to replace a channel vs. returning it to the pool, we explicitly check the channel itself for health status. If the channel is active, it will be returned to the pool, otherwise, it will be replaced.
Netty closes the channel on IOExceptions and certain other non-recoverable exceptions. The connection will be replaced as per the new logic which is the same behavior as earlier. [ Full content available at: https://github.com/apache/tinkerpop/pull/1081 ] This message was relayed via gitbox.apache.org for [email protected]
