So I’ve been a pihole user for a long long time…but seeing the advancements in AdGuard Home and some of the nicer UI facets, I was interested in giving it a try. I also have an active directory domain that I need to manage as well.

So, prior to recently, I had routed all DNS requests thought the AD DCs, and their upstream resolver was PiHole, and then Pihole routed to its internal install of cloudflared with DNS over HTTPS to the cloudflare DNS services.

More recently, I changed my DNS services in DNS to point directly to pihole, managed my local dns records in pihole and then used conditional forwarding to my AD DCs for local DNS resolution. The biggest benefit I saw in this adjustment is that I can identify what hosts are making what requests.

More recently than that, I brought Adguard Home into the environment and am using it as a secondary DNS server. I ended up taking it out of the mix for the moment. My thought process was having one DNS server on each of my active VM hosts just in case…but managing internal DNS records in adguard home is a bit of a pain in the ass, and there is no way to import in bulk.

So, the questions, 1) do you just use one or the other… pihole, vs adguard home… 2) do you use multiple dns servers or just a single one upstream…3) whats your preferred method of internal dns management in conjunction w/ pihole/adguard home?

  • Cynyr36@alien.topBEnglish
    1·
    10 months ago

    Dual unbound servers running unbound-adblock in recursive mode with DNSSEC on, with a stubzone for my internal domain (*.lan) pointed at the dnsmasq server that handles dhcp and local DNS.

    I wanted dns redundancy so at least “the Internet” would work if I was rebooting something, which the sub zone handles very well.

    Dnsmasq is set to no upstreams, and authoritative for the domain. This gives me ddns for clients as well.

    I did look into kea for DHCP and nsd for local DNS, but kea wasn’t really ready to handle dual stack clients with the ddns updates. It was neat that you can run kea in a proper redundant config. Not sure I’d have been able to get the ddns updates to dual nsd servers working without a hidden primary, leaving me with a single point failure.