Hi all,
I would like some clarifications on how the HTTP-based renewal of a certificate with multiple SAN is expected to work when some of the domain in the list are not available and/or not managed by my server anymore.
Premise: I read the thread here, but I see mixed (and unexptected) results.
Example: I obtain a certificate for www2.example.com, www3.example.com and www4.example.com via the following command: certbot certonly --webroot -w /var/www/html -d www2.example.com -d www3.example.com -d www4.example.com
When I obtained the certificate via HTTP challenge, all hostname where clearly managed by my server. After that, www3.example.com is deleted and www4.example.com is changed to point an external server that I do not control.
Now the strange part:
- if I force a renewal re-executing
certbot certonly --webroot -w /var/www/html -d www2.example.com -d www3.example.com -d www4.example.com
, it seems to work without issues even if www3.example.com and www4.example.com can’t really be renewed via the HTTP challenge; - if I use add “–dry-run” to the command above, verification fails stating “Challenge failed for domain …”, as expected (and as suggested in the linked thread).
So, my questions:
- why I see such different behavior?
- are the hostnames/domains really checked via HTTP challenge during a renewal? If so, why a forced renewal succeeds?
Thanks.