macOS DNS resolve issue in internal network with Bind9 as DNS service

pr flag

I'm using Bind9 as a DNS service to resolve some domain names internally. The topology is simple, with one router for all internal computers, and all computers using an internal DNS to resolve both internal and external domain names.

Let's say we have a domain name "", which has a public A record handled by a service provider like GoDaddy and is assigned to a public IP address, while my internal DNS is set (the same domain name) to resolve it as a private IP address internally.

The interesting thing is that on macOS (Ventura), if I use ping or a browser like Safari, that domain name will always be resolved to the public IP. However, if I use dig or nslookup, the resolver and result are correct with the private IP. Flushing DNS (using "dscacheutil -flushcache" and "sudo killall -HUP mDNSResponder", also clearing Bind9 cache) does not resolve this issue.

All these machines have IP addresses and DNS server IP allocated by a DHCP server, and that DNS is the only source for resolving all domain names.

All other systems like Linux or Windows resolve the same domain name correctly to the designated private IP without any problem. Also, it looks like only the domain (or subdomain) that has been assigned a public IP will be resolved incorrectly on macOS (all non-assigned domain or subdomain names are correct internally). So I'm guessing Bind9 is resolving (or recursively using) some forwarders?

are there any configurations to enforce resolving the internal record first? or is there something I need to fix specifically for this case?


Thank you for any help!

pr flag

Ok, problem solved. It turned out to be the iCloud+ Private Relay is routing traffic to Apple's own DNS server. I think it will probably only affect Ping and Safari. Turn it off will resolve it. Hope this will help if folks running into this issue in future.

I sit in a Tesla and translated this thread with Ai:


Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.