Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> Nice. DNS has grown notoriously complex over the years and it is hard work to run a standards compliant service. Congratulations.

Running your own private recursive resolver is very easy.

Edit: I believe people are confusing running a DNS nameserver with running a DNS resolver. The former might be hard, the latter is very easy.



No. Its not. You need to buy hardware, set it up and later constantly maintain it. It requires technical knowledge, willingness to do it and, above all, free time. While initial investment may seem trivial, on the long run it's not. People very often tend to forget that own solutions are not set up and forget. This is why cloud services are a thing in a first place.


You can absolutely just run a recursive resolver on your laptop use that resolver in every[+] network. There is literally nothing special about a recursive resolver except it doing some legwork that a stub resolver / filter resolver (like glibc or dnsmasq) doesn't do.

If you run e.g. Linux or BSD, you'd just install knot-resolver, enable the service and put "127.0.0.1" in your /etc/resolv.conf. That's it.

Similarly if you run something like pihole it is very easy to have it run a recursive resolver as well, I bet pihole has a page on how to set that up, and I doubt it is hard in any way.

If, of course, all you have now is a router provided by your ISP and you want to run your own intranet DNS resolver, then, yeah, you'll probably need some hardware for that. Obviously.

[+] some networks hijack outgoing DNS.


Debian 10 Buster already install Stubby DNS by default as your caching DNS server.


Stubby is, as the name implies, not a recursive resolver. It's a DoT stub resolver.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: