Refresh zone after record is deleted

[sauerj]

After the ACME challenge is removed from the zone, the new zone data
must be applied to the DNS servers.

[sauerj]

Sorry, I messed up this pull request. Fixes are on the way.

[sauerj]

Now everything should work.

[mcdado]

Thank you again for contributing!
Unfortunately it doesn't work yet, the async POST call to refresh the zone never happens because process.exit(0) is called before. You could simply delete that line, since it's completely redundant. Also, I personally like to use different exit codes, so I know where the process exited. You could simply go up a number and use process.exit(3) instead of process.exit(1).

[sauerj]

Hey, I am completely new to node.js...
I see, the POST call is actually inside the DELETE call. That was not what I intended to do.

About the exit code, in 'bin/create_record.js' you used process.exit(1) for exiting when a error occurred. I overlooked that you use different exit codes in 'delete_record.js'.

[mcdado]

We've all been there! ;)

Well, it is actually good the way you did it. You have to wait for the call back from the DELETE call.
The problem is that ovh.request('POST',... puts an async operation on the event loop to not block the thread, meaning it continues the flow, hitting process.exit(0) and closing the whole program all together

[mcdado]

I think simply removing process.exit(0); will be enough.

[mcdado]

Well, now this is even worse 😅
Basically what will happen is:

1. DELETE request is scheduled
2. POST request is scheduled
3. Process exited (0)
4. DELETE & POST request are executed at the same time.

Of course 4 never happens because the process is already exited. The way it was done before it was perfect. Before, the flow was:

1. DELETE is scheduled
2. DELETE is executed
3. Callback of DELETE is executed
4. POST is scheduled
5. Process is exited (0)
6. POST is executed.

The problem here was simply that the process exited before the refresh call could be executed. I suggest to go back to the previous commits, simply deleting the process.exit(0); line that is superfluous and actually bad.

[sauerj]

Okay, to be clear: you wrote the process.exit(0); not me. That actually confused me a bit.
What confuses me a little bit more, the code actually works fine. :D

[mcdado]

I know, I wrote it, but when I did there was nothing else to do… while now you have also refresh the zone — which btw is a good idea because it is a necessary step to update the zone.

In my tests it actually doesn't delete the record… you can test by running dig against the DNS record or using the API console here: https://api.ovh.com/console/

[sauerj]

Okay, I have to excuse. You are right, the record will not be deleted. It is possible that I ran the test with older code... oops.

[mcdado]

Thank you again for your contributions!