https://rancher.com/ logo
Join Slack
Powered by
more generally, i just set up a k3s cluster on 6 r...
# general
h
more generally, i just set up a k3s cluster on 6 raspberry pis and i noticed two pods in kube-system (metrics-server and local-path-provisioner) were constantly in a crashloopbackoff state so i’m trying to diagnose that error. my cluster architecture seems to be kind of odd so if anyone wants more details i can provide more.
c
what are the errors in the pod logs
h
here’s the error for local-path-provisioner
here is the error for metrics-server
there’s additional text about flags
and then this at the end
sorry this is my first experience with kubernetes in general so im a bit of a novice
c
Did you make sure you have all the correct kernel modules installed?
See the node about the vxlan kernel module here: https://docs.k3s.io/installation/requirements?os=pi#operating-systems
Also make sure you have any firewalls (ufw/firewalld) disabled on all the nodes
h
it looks like i already have the vxlan module
i’ll try disabling firewalls
the way my cluster is set up is that my control panel node is also a dns server for the worker nodes, does that affect things?
is there a way i could manually allow traffic from 10.43.0.1 through iptables?
c
We recommend just disabling the host firewall, especially if you're just starting out and are just trying to get things working.
h
thanks! i saw this, does this work?
for legacy iptables i mean
i just tried this and it looks like the pods aren’t looping and restarting anymore, thank you!
Open in Slack
PreviousNext
Powered by