After upgrading from 24.04 to 25.1 I found that I couldn’t install apps via the catalog page. After digging down I sequentially found that I couldn’t connect to Docker at all, and then digging further I discovered that my system is completely unable to perform DNS lookups via UDP. Forcing TCP allows my machine to connect. I have no idea what to even do here, I’m pretty unfamiliar with a lot of networking stuff. All I really know is that running dig <url> +tcp works and running dig <url> +notcp times out. I’m fairly confident that this is a problem on my system and not related to my router or ISP as I’m able to connect just fine on every other computer on my network. Any help is appreciated. I feel like I’ve everything short of re-installing the operating system
Hm okay I definitely missed this deadline but I guess my follow-up question would be is this related to a system-wide networking thing? When I digany URL it times out, not just docker URLs.
If this is the underlying issue then I assume the only fix would be to re-install the OS?
Post how you have your networking set up. Others will use that to make advice or try to diag the problem. Looking for things like if using DHCP or static ips with manual configuration for routing, etc. Are you using a bridge or have more advanced network setup. Is there a firewall setup?
My networking hardware setup is very basic, TP Link CR700 modem/router. No firewall on the router currently. Here’s my network config on TrueNAS. I have DHCP on and my network interface is enp0s31f6 for whatever that’s worth
Any insight on how I could check or test this? The only place I’ve seen 53 is when I run a dig command in the shell it says ; ; communications error to 8.8.8.8#53: timed out. Also when I run dig forcing TCP it is able to connect and shows SERVER: 8.8.8.8#53(8.8.8.8) (TCP)
you could try to rebind the incus port and free up port 53
This command from shell rebinds the incus port from 53 to 5354
sudo incus network set incusbr0 raw.dnsmasq="port=5354"
My thinking is that incus grabbed port 53 for itself and now truenas doesn’t have access to port 53 fort its reseolved process and can’t loockup requests
root@truenas[~]# sudo incus network set incusbr0 raw.dnsmasq="port=5354"
If this is your first time running Incus on this machine, you should also run: incus admin init
To start your first container, try: incus launch images:ubuntu/22.04
Or for a virtual machine: incus launch images:ubuntu/22.04 --vm
Error: Network not found
Here’s the interface config, I think. I don’t even know how to answer that question haha I apologize I’m fairly unfamiliar with networking stuff. I don’t think I have a bridge?
What happens if you replace the hard-coded Google/CloudFlare DNS servers with the DNS servers from your router? This smells like something local to your network.
I found this post in my initial investigation and tried your posted fix. I didn’t walk through all of your recreation steps for your original issue so not sure mine is the same, but I did create the override.conf file and that didn’t fix things. Also I didn’t already have a directory for containerd.service.d so I had to make that first; not sure if that’s a clue
I just added a workaround to force the OS to use TCP for all DNS by adding options use-vc to /etc/resolv.conf . I believe this is similar to @dain ‘s fix, but he only forced TCP for containerd rather than system-wide?
I’ve also added a post-init script to make sure this option is set on startup (ChatGPT said TrueNAS might overwrite this file whenever the system reboots).
I’m also curious if this is a safe workaround? Again my lack of understanding of all things networking has me anxious about doing anything janky like this. Some quick googling tells me that the only concern might be performance?
