This message was deleted.
# k3sa
adamant-kite-43734
11/11/2023, 2:09 PMThis message was deleted.
👍 1
b
brave-afternoon-4801
11/11/2023, 2:10 PMHave you tried launching one-off pods, shelling in, and testing out dns? If not, I would probably try that next
a
astonishing-train-67743
11/11/2023, 2:11 PMb
brave-afternoon-4801
11/11/2023, 2:11 PMIt's also possible that your nodes have a messed up /etc/resolve.conf
brave-afternoon-4801
11/11/2023, 2:11 PMhmm
brave-afternoon-4801
11/11/2023, 2:11 PMwell, if coredns isn't running, it kinda makes sense that cluster level dns wouldn't be around, but i don't know how you bootstrap in that case
a
astonishing-train-67743
11/11/2023, 2:12 PMif i append the coredns service ip to the end of the command it works if not it returns recursion not available from 10.43.0.10 (which is coredns svc ip)
b
brave-afternoon-4801
11/11/2023, 2:12 PMhm interesting
a
astonishing-train-67743
11/11/2023, 2:15 PMjust signed up to CNCF slack to ask in the CoreDNS channel aswell, I added the resolvconf flag to systemd and restarted to make it use a different resolv file instead of the systemd-resolved cache Fedora runs
b
brave-afternoon-4801
11/11/2023, 2:16 PMi hope you get an answer. it's an interesting question for sure 🙂
a
astonishing-train-67743
11/11/2023, 2:17 PMme to
astonishing-train-67743
11/11/2023, 2:19 PMLooking at coredns logs it seems to be adding an additional namespace to the failing requests
b
brave-afternoon-4801
11/14/2023, 12:24 PMI ended up making a vagrant box to replicate your problem - but I can't. coredns always comes up smooth.
brave-afternoon-4801
11/14/2023, 12:25 PMhowever, cilium is not completely happy, so 😞
brave-afternoon-4801
11/20/2023, 7:44 PMWell, now I don't know what problem you had. I got cilium working well now, even as the ingress controller