I ran this command: ./certbot-auto certonly --webroot -w /var/www/html/wawl -d wawl.org
It produced this output: Failed authorization procedure. wawl.org (http-01): urn:ietf:params:acme:error:dns :: DNS problem: SERVFAIL looking up A for wawl.org
My web server is (include version): Apache 2.2.15
The operating system my web server runs on is (include version): RHEL 6.9
My hosting provider, if applicable, is: Self-hosted
I can login to a root shell on my machine (yes or no, or I don’t know): yes
I’m using a control panel to manage my site (no, or provide the name and version of the control panel): no
Do you know the wawl.org or chattanoogastate.edu domain and DNS admins? They could make changes to how one or both of the domains are set up to avoid this. For example, you could:
Delete some of chattanoogastate.edu's DS records.
Move wawl.org to different nameservers. For example, create ns1.wawl.org and ns2.wawl.org using the current two IPs. You could also move to a completely different DNS service, but that’s obviously a lot of work.
Let’s Encrypt could change their DNS resolver configuration, but it’s set up this way for security reasons, and they probably won’t.
Long term…
I’ve asked some DNS people what they think about it.
Verisign (the .edu TLD operator) could change their DNS server to handle this differently (like by setting the TC bit), but we’d have to read the specifications and think about it before drawing any conclusions.
Unbound (the DNS resolver software Let’s Encrypt uses) could be modified to handle this situation differently (like by automatically falling back to TCP), but I don’t have an opinion on whether it should be.
I don’t have access to and would not touch these things as they are not in my knowledge realm. Not that I don’t have a passing knowledge, but my focus is web development.
Our network team will need to look at this and may even say “no, we’re not making changes” in which case I will have to go a different route for my certificate.
You are right on the mark. College IT systems can be weird!
Edit: By the way, you could also “fix” the problem by making the DS and/or NS record sets bigger. If the response was 13 bytes larger, the authoritative DNS servers would either set the truncation bit, or remove the DS and RRSIG records, allowing the NS and A records to fit. This would be gross, less efficient, and might result in resolution issues with a small percentage of clients.