Closed SchroterQuentin closed 3 years ago
Why do you have default routes on both interfaces? Does configuration of the additional interface still cause issues if you don't add a second default route?
I removed all the routes
and routing-policy
configuration and the issue is gone.
My bad
K3s Version: k3s version v1.18.9+k3s1 (630bebf9)
Fresh new Ubuntu 18.04 virtual machine on vmware with 2 vCPU
Cluster Configuration: Single node
Describe the bug:
When adding a new network interface on the virtual machine, pods CoreDNS become unready and provisioner storage restart, both seems to failed because of a network issue.
Here is the netplan :
And if you execute this two commands, the issue is gone :
I found some issue about flannel which is using the first network interface when installing by k3s but I don't know much more in this area.