-
Bug
-
Resolution: Unresolved
-
Undefined
-
None
-
4.18
-
No
-
False
-
Description of problem:
Create layer 2 udn in one ns and enable multicast on this ns, create testing pods, and send multicast traffic on all pods to join a multicast group. Check on pods, multicast traffic is not received.
Version-Release number of selected component (if applicable):
built image via 'build 4.18.0-0.nightly,openshift/api#2127,openshift/ovn-kubernetes#2413'
How reproducible:
always
Steps to Reproduce:
Steps to Reproduce:
1. create ns test and enable multicast
oc annotate namespace test k8s.ovn.org/multicast-enabled=true
2. create layer 2 udn in ns test
% cat udn_l2.yaml
apiVersion: k8s.ovn.org/v1
kind: UserDefinedNetwork
metadata:
name: udn-l2
spec:
topology: Layer2
layer2:
role: Primary
subnets:
- "10.0.0.0/24"
3. create testing pods
% cat multicast-rc.json
{
"apiVersion": "v1",
"kind": "ReplicationController",
"metadata": {
"labels": {
"name": "mcast-rc"
},
"name": "mcast-rc"
},
"spec": {
"replicas": 3,
"template": {
"metadata": {
"labels": {
"name": "mcast-pods"
}
},
"spec": {
"containers": [
{
"image": "quay.io/openshifttest/mcast-pod@sha256:aa2a8fdcfeddb09097e6a7201cfdb062db8d92d523cbcf758ef7d3545a1f5776",
"name": "mcast-pod"
}
]
}
}
}
}
4. sending multicast traffic on all pods, the ip list is the udn ips of 3 pods
e.g: % oc rsh mcast-rc-6s2qd ~ $ omping -c 5 -T 15 10.100.3.3 10.100.2.5 10.100.1.3
5. find out that pods can not receive multicast traffic
10.151.1.3 : unicast, xmt/rcv/%loss = 10/10/0%, min/avg/max/std-dev = 1.162/1.250/1.348/0.055 10.151.1.3 : multicast, xmt/rcv/%loss = 10/10/0%, min/avg/max/std-dev = 1.214/1.336/1.921/0.209 10.151.5.3 : unicast, xmt/rcv/%loss = 10/10/0%, min/avg/max/std-dev = 0.454/0.554/0.623/0.046 10.151.5.3 : multicast, xmt/rcv/%loss = 10/10/0%, min/avg/max/std-dev = 0.468/0.568/0.641/0.047
Actual results:
multicast traffic is dropped
Expected results:
multicast traffic should be received.