In the cluster, replication and sentinel modes, the first thing we do is iterate through known nodes and for each, connect and send a command to collect some information, until we obtain a successful response.
This commit makes sure we remember the failures and when no successful response is ever obtained, we can complete the connection Future with an error message that contains the actual failures and not just a placeholder string.
In the cluster, replication and sentinel modes, the first thing we do is iterate through known nodes and for each, connect and send a command to collect some information, until we obtain a successful response.
This commit makes sure we remember the failures and when no successful response is ever obtained, we can complete the connection
Future
with an error message that contains the actual failures and not just a placeholder string.Backport of #426 to 4.x