After the DNS challenge fails, each order creation will not be reused

The process is as follows: When applying for a certificate via DNS Challenge, if the client fails to complete the challenge for some reason, the client sets the order to "deactivated". The client then initiates a second application, obtaining a new Authorization URL. However, instead of proceeding with the challenge, we directly initiate a third order creation, at which point we receive an Authorization URL that differs from the second one. After a successful challenge, if we retry the steps above, the returned results become as expected.

If after initially creating an order, we create a new order without attempting the challenge, the Authorization URL we obtain remains unchanged.

Although this doesn't cause any actual issues, it seems inconsistent with our expected behavior.

We set up a local ACME CA Server using GitHub - letsencrypt/boulder: An ACME-based certificate authority, written in Go. for testing, but after creating the third order, the Authorization URL we retrieve matches the one from the second attempt.

What differences might exist in the production environment?

1 Like

In your test Boulder instance, do you have the feature flag NoPendingAuthzReuse: true? I believe that’s set for the Let’s Encrypt instance prod instance of Boulder, and causes the behaviour change you’re describing.

4 Likes

I am using docker-compose to start instances, and I see that NoPendingAuthzReuse: true in test/config-next/ra.json.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.