arp table do not be updated from source pod when destination pod mac is changed but with same ip

    4.16.0-0.nightly-2024-03-06-073110
    sriov-network-operator.v4.16.0-202403050942

    always

    1. setup sriov operator
    2. Create 4 VF by creating snnp 

apiVersion: sriovnetwork.openshift.io/v1
kind: SriovNetworkNodePolicy
metadata:
  name: e810c1
  namespace: openshift-sriov-network-operator
spec:
  deviceType: netdevice
  nicSelector:
    deviceID: "1593"
    pfNames:
      - ens2f2
    vendor: '8086'
  nodeSelector:
    feature.node.kubernetes.io/sriov-capable: 'true'
  numVfs: 4
  priority: 99
  resourceName: e810c
        
    3. create namespace z1 and sriovnetwork cr which only have 2 ips can be used. 

apiVersion: sriovnetwork.openshift.io/v1
kind: SriovNetwork
metadata:
  name: sriovnetwork
  namespace: openshift-sriov-network-operator
spec:
  ipam: |
    {
      "type": "whereabouts",
      "range":"10.30.0.0/30"
    }
  capabilities: |
    {
      "mac": true,
      "ips": true
    }
  spoofChk: "off"
  trust: "on"
  resourceName: e810c
  networkNamespace: z1

4. Check the NAD on z1
# oc get net-attach-def -n z1
NAME           AGE
sriovnetwork   61m

5. Create 2 test pods with the following, there will be 2 pods running

apiVersion: v1
kind: Pod
metadata:
  generateName: testpod1
  namespace: z1
  labels:
    env: test
  annotations:
    k8s.v1.cni.cncf.io/networks: sriovnetwork
spec:
  securityContext:
    runAsNonRoot: true
    seccompProfile:
      type: RuntimeDefault
  containers:
  - name: test-pod
    image: quay.io/openshifttest/hello-sdn@sha256:c89445416459e7adea9a5a416b3365ed3d74f2491beb904d61dc8d1eb89a72a4
    securityContext:
      allowPrivilegeEscalation: false
      capabilities:
        drop: [ALL]
  automountServiceAccountToken: false
     

5. Create another test pod again. the pod will be in ContainerCreating since no ip can be used

6. ssh into one pod and ping another running pod and then check the arp table 

 # oc rsh -n z1 testpod1mjq7w
~ $ arp -a
? (10.131.2.1) at 0a:58:0a:83:02:01 [ether] on eth0
? (10.30.0.1) at ce:56:74:32:f6:d1 [ether] on net1

7. Delete the pod with ip 10.30.0.1, then a new pod running with ip 10.30.0.1 but mac address is changed

# oc exec -n z1 testpod186fxn ip a s net1
kubectl exec [POD] [COMMAND] is DEPRECATED and will be removed in a future version. Use kubectl exec [POD] -- [COMMAND] instead.
45: net1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether 82:bc:7e:ca:92:bd brd ff:ff:ff:ff:ff:ff
    altname enp97s0f2v0
    inet 10.30.0.1/30 brd 10.30.0.3 scope global net1
       valid_lft forever preferred_lft forever
    inet6 fe80::80bc:7eff:feca:92bd/64 scope link 
       valid_lft forever preferred_lft forever

8. Check the arp again of step 6, the mac address still old. 

# oc rsh -n z1 testpod1mjq7w
~ $ arp -a
? (10.131.2.1) at 0a:58:0a:83:02:01 [ether] on eth0
? (10.30.0.1) at ce:56:74:32:f6:d1 [ether] on net1

arp table mac address still old pod mac    

(10.30.0.1) at ce:56:74:32:f6:d1 [ether] on net1

 arp table mac address should be updated to
   
 (10.30.0.1) at 82:bc:7e:ca:92:bd [ether] on net1