bpfin.cali bpfout.cali 1500 MTU while OS eth has MTU 9000 and calico vxlan interfaces have MTU 8950 is it affects performance?

[AleksandrSor]

Expected Behavior

Be able to increase MTU if it affects performance

Current Behavior

I am not sure if we have a problem here

Possible Solution

Steps to Reproduce (for bugs)

1. setup the cluster with configs below
2. verify net connections

Context

bond0: connected to bond0
        "bond0"
        bond, B8:CE:F6:CC:A3:F4, sw, mtu 9000
        ip4 default
        inet4 172.24.1.24/24
        route4 default via 172.24.1.1 metric 300
        route4 172.24.1.0/24 metric 300

ens1f0np0: connected to ens1f0np0
        "Mellanox MT2892"
        ethernet (mlx5_core), B8:CE:F6:CC:A3:F4, hw, mtu 9000
        master bond0

ens1f1np1: connected to ens1f1np1
        "Mellanox MT2892"
        ethernet (mlx5_core), B8:CE:F6:CC:A3:F5, hw, mtu 9000
        master bond0

lo: connected (externally) to lo
        "lo"
        loopback (unknown), 00:00:00:00:00:00, sw, mtu 65536
        inet4 127.0.0.1/8

eno8303: disconnected
        "Broadcom and subsidiaries NetXtreme BCM5720"
        ethernet (tg3), 70:B5:E8:D0:A0:F0, hw, mtu 1500

eno8403: disconnected
        "Broadcom and subsidiaries NetXtreme BCM5720"
        ethernet (tg3), 70:B5:E8:D0:A0:F1, hw, mtu 1500

bpfin.cali: unmanaged
        "bpfin.cali"
        ethernet (veth), 7E:D6:B4:AE:4A:2E, sw, mtu 1500

bpfout.cali: unmanaged
        "bpfout.cali"
        ethernet (veth), 46:D9:B4:BA:E4:2F, sw, mtu 1500

cali00ba82360c7: unmanaged
        "cali00ba82360c7"
        ethernet (veth), EE:EE:EE:EE:EE:EE, sw, mtu 8950

cali010ee92ee2b: unmanaged
        "cali010ee92ee2b"
        ethernet (veth), EE:EE:EE:EE:EE:EE, sw, mtu 8950

Your Environment

• Calico version: v3.27.3 v3.28.0
• Helm chart: projectcalico/tigera-operator:v3.28.0

  - chart: projectcalico/tigera-operator
    version: v3.28.0
    name: calico 
    namespace: tigera-operator
    values:
      - installation:
          calicoNetwork:
            linuxDataplane: BPF
            mtu: 8950
            bgp: Disabled
            ipPools:
              - cidr: 10.244.0.0/14
                blockSize: 20
                encapsulation: VXLAN

kind: ConfigMap
apiVersion: v1
metadata:
  name: kubernetes-services-endpoint
  namespace: tigera-operator
data:
  KUBERNETES_SERVICE_HOST: "172.24.1.15"
  KUBERNETES_SERVICE_PORT: "6443"

apiVersion: projectcalico.org/v3
kind: FelixConfiguration
metadata:
  annotations:
    operator.tigera.io/bpfEnabled: "true"
  creationTimestamp: "2023-11-09T09:16:01Z"
  generation: 1
  name: default
  namespace: calico-system
  resourceVersion: "65934135"
  uid: 5d0de475-8603-4e7b-9282-baecc231e48e
spec:
  bpfEnabled: true
  bpfExternalServiceMode: DSR
  bpfLogLevel: ""
  floatingIPs: Disabled
  healthPort: 9099
  logSeverityScreen: Info
  reportingInterval: 0s
  vxlanMTU: 8950
  vxlanVNI: 4096

# /etc/NetworkManager/conf.d/calico.conf
[keyfile]
unmanaged-devices=interface-name:cali*;interface-name:bpf*.cali*;interface-name:tunl*;interface-name:vxlan.calico;interface-name:vxlan-v6.calico;interface-name:wireguard.cali;interface-name:wg-v6.cali

apiVersion: kubeadm.k8s.io/v1beta3
kind: InitConfiguration
certificateKey: ""
skipPhases:
  - addon/kube-proxy
---
apiVersion: kubeadm.k8s.io/v1beta3
kind: ClusterConfiguration
networking:
  serviceSubnet: "10.243.0.0/16"
  podSubnet: "10.244.0.0/14"
  dnsDomain: "l8s.local"
controllerManager:
  extraArgs:
    "node-cidr-mask-size": "20"
    "allocate-node-cidrs": "false"
apiServer:
  certSANs:
    - "172.24.1.15"
    - "172.24.1.16"
    - "k8clust-lon01.l8s.space"
clusterName: "k8clust-lon01"
controlPlaneEndpoint: "k8clust-lon01.l8s.space:6443"
---
apiVersion: kubelet.config.k8s.io/v1beta1
kind: KubeletConfiguration
maxPods: 4000
cgroupDriver: systemd
serverTLSBootstrap: true

• Orchestrator version:
  kubernetes 1.29.5
• Operating System and version:

# uname -r
5.14.0-452.el9.x86_64
# cat /etc/os-release
NAME="CentOS Stream"
VERSION="9"
ID="centos"
ID_LIKE="rhel fedora"
VERSION_ID="9"
PLATFORM_ID="platform:el9"
PRETTY_NAME="CentOS Stream 9"
ANSI_COLOR="0;31"
LOGO="fedora-logo-icon"
CPE_NAME="cpe:/o:centos:centos:9"
HOME_URL="https://centos.org/"
BUG_REPORT_URL="https://issues.redhat.com/"
REDHAT_SUPPORT_PRODUCT="Red Hat Enterprise Linux 9"
REDHAT_SUPPORT_PRODUCT_VERSION="CentOS Stream"

[tomastigera]

yeah that sounds like the device should have a large MTU, thanks for reporting it!