I just struggled for over one hour with a configuration collapse that appears to have something to do with IPv6 option. The setup, which had been working perfectly for the last year:
Mac Mini M1 running HQP Desktop, set to connect to NAA IPv6
UP Gateway running NAA image
Holo Spring 2 KTE
Roon on a Ubuntu Server PC
UniFi network controlled by a UniFi Dream Machine
This setup was working yesterday. Then, thanks to the high winds we had today on the Northern California coast that cause tree branches to hit power lines, we had two short power outages. We have an automatic backup generator, but it takes a few seconds to start, and in the meanwhile all computers and network gear lose power and reboot. Everything seemed to have got back up, but when I tried to play to HQP via Roon, I got an error message that HQP was not responding, even though I could see HQP Desktop up and running on the Mac mini. After a lot of messing around, including installing the latest NAA image, rebootings, and logfile inspection, I concluded that the the HQP IPv6 option, which had been working fine for the last year, seemed to be the problem. After turning that off and going through another round of rebootings, music again!
I’m puzzled why the IPv6 option worked for so long and now fails. As far as I can see, my network/router configurations have been unchanged since I set it all up November 2021, and there haven’t been any updates for the last few months.
Maybe some update got installed as result of reboot? Either in networking gear or somewhere else?
Another possibility is that devices came up in different order than before.
On Mac Mini, if you don’t use WiFi, please check that it is turned off. MacOS is a bit tricky with IPv6 at times, but IPv6 is nice in a way that even if network infra doesn’t support it, the devices still auto-configure common local network without IP conflicts. This can happen because the 48 bit MAC can completely fit in 64-bit IPv6 address.
Mac Mini WiFi is off, has been since the beginning. As far as I can see, nothing got updated, it would have taken a lot longer than what I observed when power came back (the glitch between power out and generator kicking in is around 30 seconds). I’m stumped.
Seems like some hiccup in macOS. Please check the lines under heading “NAA output network interfaces” in the log. It should list just two items and both with both IPv4 and IPv6. First likely “en0” which is your wired Ethernet, and second “lo0” which is the local loopback interface.
For some reason it gives “No route” when sending packet to a multicast address. I guess there’s an IPv4 default route on en0 for internet access. But maybe there’s no suitable route in routing table for IPv6.
“netstat -r -n” will show all routes, but it is pretty lengthy. There should be pretty universal fe80 routes on the link. Maybe also “route -n monitor” and then try IPv6 discovery would show something?
And here’s what route -n monitor says when I try to turn on IPv6 NAA:
got message of size 68 on Wed Feb 22 19:59:58 2023
RTM_DELMADDR: multicast group membership removed from iface: len 68,
sockaddrs: <GATEWAY,IFP,IFA>
1.0.5e.40.0.c7 en0:14.98.77.83.13.be 239.192.0.199
got message of size 68 on Wed Feb 22 20:00:02 2023
RTM_NEWMADDR: new multicast group membership on iface: len 68,
sockaddrs: <GATEWAY,IFP,IFA>
1.0.5e.40.0.c7 en0:14.98.77.83.13.be 239.192.0.199
And there is only IPv4, but no IPv6 (HQPlayer is dual-stack)… I’m sort of out of ideas right now why the IPv6 message send returns “no route” and doesn’t go anywhere…