Kubeadm Token Create Fails with 'route ip+net: no such network interface

DevOps Kubernetes
1

I am trying to print the token with kubeadm, but on some control plane nodes in a specific cluster I get this error:

I0122 16:14:35.529340     433 token.go:119] [token] validating mixed arguments
I0122 16:14:35.529409     433 token.go:128] [token] getting Clientsets from kubeconfig file
I0122 16:14:35.529434     433 cmdutil.go:93] Using kubeconfig file: /etc/kubernetes/admin.conf
I0122 16:14:35.530344     433 loader.go:395] Config loaded from file:  /etc/kubernetes/admin.conf
I0122 16:14:35.531511     433 token.go:243] [token] loading configurations
I0122 16:14:35.531650     433 initconfiguration.go:114] skip CRI socket detection, fill with the default CRI socket unix:///var/run/containerd/containerd.sock
I0122 16:18:00.849123     433 interface.go:432] Looking for default routes with IPv4 addresses
I0122 16:18:00.849164     433 interface.go:437] Default route transits interface "*"
route ip+net: no such network interface
[16:18] server123.place10:~# ```


Does anyone know what the real reason for this is? the node is reachable and working just fine, even though the error message seems to point in the direction of a network problem
2

kubeadm is trying to chooseHostInterface for the node, it’ll traverse the interfaces in both files:

> cat /proc/net/route
Iface	Destination	Gateway 	Flags	RefCnt	Use	Metric	Mask		MTU	Window	IRTT
eth0	00000000	FDEF1AAC	0003	0	0	0	00000000	0	0	0
cni0	0000F40A	00000000	0001	0	0	0	00FFFFFF	0	0	0
eth0	0000FEA9	00000000	0001	0	0	1002	0000FFFF	0	0	0
docker0	000011AC	00000000	0001	0	0	0	0000FFFF	0	0	0
eth0	00E01AAC	00000000	0001	0	0	0	00F0FFFF	0	0	0
root@k8s:pts/0->/root (0)
> cat /proc/net/ipv6_route```


Then try to get the the interface By callin syscall.NetlinkRIB(syscall.RTM_GETLINK, <http://syscall.AF|syscall.AF>_UNSPEC), corresponding commands is :ip link show. The problem is that it can’t find an interface name “*” from the interfaceTable