This is an usability issue. If I pass an invalid value to --log-to, I get a warning only after takeover is already entering stage 2, so I can't safely stop the process and fix my mistake.
Ideally, the validity of the argument should be checked earlier on. And I guess that by default (unless we also pass --no-ack?) we should stop the execution if the argument is not valid.
This is an usability issue. If I pass an invalid value to
--log-to
, I get a warning only aftertakeover
is already entering stage 2, so I can't safely stop the process and fix my mistake.Ideally, the validity of the argument should be checked earlier on. And I guess that by default (unless we also pass
--no-ack
?) we should stop the execution if the argument is not valid.