Been tearing my hair out recently, basically ever since I migrated my home server from Debian 11 to Ubuntu Server 24.0.4.2 LTS, and (after regretting the jump to Ubuntu) then back to Debian again, this time version 12 stable. Prior to this, just about everything relating to IPv6 on my home network was working flawlessly. I could connect to devices using either IP from either their public or private prefix that I have set up in my network, or even just DNS a majority of the time, with both A and AAAA records in place. And if something wasn't connecting, it was usually a case of a service not running (e.g. "connection refused"), or a port being blocked by one of the firewalls.
Around the same time I migrated away from Debian 11, I did change some network config on my home router. But my memory is iffy, now, on what it could have been that broke half of my IPv6 setup. Plus it was also a while ago now.
If I run a test to isp.test-ipv6.com, while on a remote desktop connection to my server in question, the test completely passes with flying colours; there I get a result code of 46 which indicates I am running a dual-stack network, as expected. So everything should be all good, right?
Well... No. Issues become apparent though when I try to connect to my home Minecraft server. IPv4 get through just fine; my router sees the incoming traffic, sees it is permitted, and passes it from the main home VLAN, into the servers VLAN. Then when it reaches my home server, the server's firewall will also permit it since it's Minecraft traffic.
If I switch things around, and try to connect with IPv6, all kinds of weird things that I cannot pin down start to arise. The main one is that a lot of traffic seems to be getting blackholed? Because if I do a packet capture while I say, connect to my Minecraft server, I see various ICMP "destination unreachable" packets being logged, which I don't see at all for IPv4 connections. ICMP echo traffic also seems really unpredictable on IPv6. I've had times where pings from my PC in my main VLAN will get through just fine to my server. But as I'm typing this, I'm trying again now, and ICMP echo requests are just timing out on me. Traceroutes aren't any better, usually timing out over and over until the max hop limit is reached.
All I want is perfect end-to-end connectivity for anyone fortunate enough to also have IPv6, who may also want to access services I am hosting over v6. And I don't have that anymore. I also hate that it's even reached this point where I even have to ask for help online, instead of being able to figure this out myself...
I would also share a saved, sample packet capture in .pcap form, but I am unsure on what the best way to do that would be, here.
Device info:
root@apollo:~# ubus call system board
{
"kernel": "5.15.167",
"hostname": "apollo",
"system": "Intel(R) Celeron(R) CPU J3455E @ 1.50GHz",
"model": "Qotom Q710G4",
"board_name": "Qotom Q710G4",
"rootfs_type": "squashfs",
"release": {
"distribution": "OpenWrt",
"version": "23.05.5",
"revision": "r24106-10cc5fcd00",
"target": "x86/64",
"description": "OpenWrt 23.05.5 r24106-10cc5fcd00"
}
}