I tried configuring the docker daemon.json like Jian A suggested, but the strange thing is that it broke ipv6 on the host system for me until I rebooted. Would like some attention on a working stable docker ipv6 config that can persist between reboots.
Edit: just wanted to include that I’m on 24.10-RC.2
Adding my voice here. I was badly surprised by this regression on upgrading to 24.10.0, where it seemed to work fine in 24.04.
I didn’t realize that one of my services was now unable to communicate with the MinIO container running on TrueNAS. The NAS itself has both IPv4 and IPv6 IP addresses on my LAN, and so I had defined it in DNS similarly. The server was selecting the IPv6 address to talk to the NAS and was silently failing until I noticed the next morning that all the media jobs had failed. Putting the IPv4 address directly into the configuration of the server is a quick work-around.
(I only upgrade TrueNAS because it addressed another bug I had in that train.)
After reboot, Truenas uses DHCPv6 gets correct IPv4 and IPv6 Adresses.
Docker daemon.json is overwritten and has to be modified to support the ipv6 subnet.
Then all works. Truenas (shell) as well as Docker Container can use IPv6 for local and global (internet) communication.
But then After some time (idk what exactly) Truenas looses its Address, Gateway and Route assigned by DHCPv6. But NO OTHER DEVICE in my network does this.
Therefore Docker has no IPv6 internet connectivity.
After this Bug? accours nothing helps than restarting truenas. Maybe there is a way to refresh the network adapter i dont see.
I have the feeling its a duality of network adapter bug and docker configuration.
I thought ixSystems were better with IPv6 …
And yes under Apps → Settings → Address Pools – its not possible add IPv6 Subnets …
