Dimitri Fontaine

Given my reading of the code I believe you've spotted a real problem with this attempt: ``` c args[argsIndex++] = primaryConnInfo; ... /* log the exact command line we're using...

I just pushed a better version of this fix, using `unsetenv`, or `setenv` again to the previous value of the PGPASSWORD environment variable if it happened to be already set...

See also #334, which I think we should pay attention to while fixing this one.

Hi @jnehlmeier ; thanks for your detailed bug report: I believe the error is found here: https://github.com/citusdata/pg_auto_failover/blob/0fecd653f9ef43019e4d76e347afce378adcc481/src/bin/pg_autoctl/pgctl.c#L1207-L1210 The way we set the replication password when using either `pg_basebackup` or `pg_rewind`...

On Mac I just install https://postgresapp.com and then the Xcode command line tools, and that should be enough to get started. I must admit I have a full dev environment...

Hi @soumyadeep2007, thanks for looking into this area in details. As you can see in https://github.com/citusdata/pg_auto_failover/blob/master/tests/test_auth.py at the moment the setting of the password is not part of the `pg_autoctl`...

Hi @soumyadeep2007, I think things are more complex than they look at first. The `--auth` option we provide is used to open the HBA for three different users: the monitor...

Something we had in the very early prototype was a `pg_autoctl create setup` command that would only check things and create a configuration, but refrain from registering to the monitor...

Hi Tiago! The idea of executing a command when there is a state change has been discussed in the past already. See https://github.com/citusdata/pg_auto_failover/issues/440 and https://github.com/citusdata/pg_auto_failover/issues/490. I think we should include...

The trick with LISTEN/NOTIFY is that we would have to execute the command each time `pg_autoctl` is started and then each time there is a failover, or say a state...