What's the file that http-01 challenge requires at challenge?


I’m using http-01 challenge. To better understand what I’m doing, I want to know what is the file that certbot (using webroot mode) puts inside .well-known/acme-challenge/TOKEN to complete the challenge.

Can you help me?

Thank you!

Hi @matheuslc

The HTTP-01 challenge mechanics are defined as part of the ACME protocol. If you want the really nitty gritty details you’d be interested in reading ACME draft-07, Section 8.3. That describes the overall HTTP-01 challenge.

The contents of the file itself are an ACME key authorization which Section 8.1 describes. Key authorizations are used with the other challenge types as well.

The condensed explanation is that Certbot creates an ACME account & keypair for you when you run it initially and register an account. The public key of that account keypair gets registered with the ACME server (Let’s Encrypt in this case).

When it comes time to prove ownership of a domain Certbot tells the server it wants to be authorized for the domain. The server sends back a pending authorization that has a HTTP-01 challenge as one option for completing the authorization. That HTTP-01 challenge comes with a random token and Certbot takes the token and uses your ACME account private key and the token to creates a key authorization string as described in Section 8.1. The key authorization string is then put in a file identified by the random token and then placed in .well-known/acme-challenge/ for the ACME server to check.

When the ACME server sends a HTTP request to yourdomain/.well-known/acme-challenge/TOKEN it can look at the key authorization string from inside that file and validate it with the token and the public key known for your account. If it checks out then your ACME account is considered authorized to issue certificates for that domain.

Hope that helps explain somewhat!


A simpler way to put this might be to say that the certificate authority chooses each time what the file name and contents should be. There is no way to predict them ahead of time because they are randomly chosen by the CA. And there’s no way to re-use one from an old challenge for a new challenge¹.

The ACME client, like Certbot or acme.sh or whatever other software you’re using, is told each time that the CA now expects a particular file at a particular place. The client will then have to respond by creating the requested file, which in turn proves to the CA that whoever is running the client is someone who actually has the ability to update /.well-known/acme-challenge on that server.

¹ Sort of. See also https://github.com/Neilpang/acme.sh/wiki/Stateless-Mode for something like an exception to this.


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