On Mon, 21 Sep 2026 18:37:04 +0200 Nicolas George <[email protected]> wrote:
> Joe (HE12026-09-21): > > For completeness, I was also in the thread early with a need to > > disable it. > > > > Short version: my ISP does not (yet) provide IPv6 connectivity, and > > my router was supplied by the ISP so already had IPv6 disabled. I > > was having trouble with DNS failures due to bind9 not finding an > > IPv6 route out of the network and giving up, even though the > > requested host also had an IPv4 address. > > Millions of people in the world use Linux with ISPs and routeurs > without IPv6 support and they do not need to disable anything. > > Your diagnostics of the problem was wrong. I'm sure you're ready to correct me. So how do I tell bind to try both protocols if it finds no route for one of them? I would quite honestly have expected it to do that automatically. It was a recent bind9 installation, with just the local zone information added. > > > I was advised that disabling IPv6 in the kernel was the final fix, > > That was wrong. I'm always willing to learn. > > > so I did that to the two Raspberry Pis which were having the > > problems, mostly of finding RPi apt repositories. No good. I had to > > disable it individually in both bind9 and apt to get reliable IPv4 > > lookups. I left the thread at this point. > > That was wrong too, but at least a wrong that actually fixed your > issue without introducing new ones that affected you. > > So *is* there a real solution, which gets me reliable DNS function without causing further problems? At the moment, it seems to me that bind9 was doing something wrong, but I'm more than willing to find that it wasn't, and just needed the right configuration. I did spend a fair amount of time trying to find what that might be before I came here, even going as far as RTFMing. I also think that the network should have functioned correctly without an external IPv6 connection, as it did with the HP microserver before that was replaced by the RPis. The RPis had lived in the network for quite some time before they took over server duties, so I had been using their repositories without any DNS troubles. There is always the unknowable possibility that bind9 running on ARM is not quite the same as bind9 running on amd64. I used to dabble with ARM assembler about thirty years ago, and I'm aware of the alien nature of the instruction set when compared to that of an Intel-style CISC processor. I very much doubt that a complex set of source code can be just dropped onto such a different compiler without some changes, or at least optimisations. There is also the possibility of the RPi networking having a problem, but there are a lot of them about and it would have surfaced before now. Not many may be running bind9, though. -- Joe
