MTU is subtracted twice: from kube-bridge and tun- interface

[zerkms]

What happened?
As it is implemented now: with auto MTU enabled - kube-bridge is configured with -20 of the physical interface. Then tun- ip-ip interfaces are configured with further -20 of the kube-bridge.

So we end up having kube-bridge with 1480, and tun- with 1460.

I believe kube-bridge's MTU should simply match the physical link MTU.

What did you expect to happen?
I think kube-bridge should always (?) match the physical link MTU.

How can we reproduce the behavior you experienced?
Steps to reproduce the behavior:

1. Deploy kube-router v1.1.1 to the aws cluster (with at least one node in different AZ)
2. Check MTUs for kube-bridge and tun-

**Screenshots / Architecture Diagrams / Network Topologies **
If applicable, add those here to help explain your problem.

** System Information (please complete the following information):**

• Kube-Router Version (kube-router --version): v1.1.1
• Kube-Router Parameters: [e.g. --run-router --run-service-proxy --enable-overlay --overlay-type=full etc.] all default
• Kubernetes Version (kubectl version) : [e.g. 1.18.3] 1.19.5
• Cloud Type: [e.g. AWS, GCP, Azure, on premise] aws
• Kubernetes Deployment Type: [e.g. EKS, GKE, Kops, Kubeadm, etc.] kubeadm
• Kube-Router Deployment Type: [e.g. DaemonSet, System Service] daemonset
• Cluster Size: [e.g. 200 Nodes] 10

** Logs, other output, metrics **
Please provide logs, other kind of output or observed metrics here.

Additional context
Add any other context about the problem here.

[zerkms]

UPD:

what's interesting, if I disable auto-mtu and set 1500 for kube-bridge manually, it still creates the following tunnel interfaces:

6: tunl0@NONE: <NOARP> mtu 1480 qdisc noop state DOWN group default qlen 1000
7: tun-10703247@ens5: <POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1460 qdisc noqueue state UNKNOWN group default qlen 1000
8: tun-10702180@ens5: <POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1460 qdisc noqueue state UNKNOWN group default qlen 1000
9: tun-10702199@ens5: <POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1460 qdisc noqueue state UNKNOWN group default qlen 1000
10: tun-10703135@ens5: <POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1460 qdisc noqueue state UNKNOWN group default qlen 1000

[murali-reddy]

Not sure why I felt MTU has to be adjusted explicitly on IP-IP tunnel interface in this PR.

#109

Tunnel intefaces MTU are automatically deducted with 20 bytes to accomidate IP-IP tunnel header.

[zerkms]

@murali-reddy does it mean db1bd56#diff-0c7a66931bb103d41704081acb720a2e87222dbd9732538d0bd911cb3106e9bcR136 this very change return lintMTU - 20, nil should not exist either? eg: kube-bridge should be auto-configured to have MTU matching the physical interface, regardless if ipip