Linux bridge ports are not reflecting the desired state for vlan IDs #631
Comments
|
Hi, vlan filtering is not supported at kubernetes-nmstate yet, we just do a best effort [1] to bridge-cni vlan feature [2] to work fine. [1] https://github.com/nmstate/kubernetes-nmstate/blob/master/build/bin/vlan-filtering |
|
@qinqon thanks for your prompt response. |
|
The nmstate is fine issue is at kubernetes-nmstate since we try to do vlan filtering on the whole range so whatever is configured with net-attach-def + bridge CNI will work fine. |
|
aha i see, i saw that |
|
Our intention was to configure all Linux bridges in a way that they treat their Now when There is a tracker for it, in case somebody would want to contribute the feature. I'd be glad to guide the effort: #171
That is what should happen automatically - all ports assigned to a bridge via the desired state should become trunks. |
|
Closing in favor of #634 |
What happened:
Create a Linux vlan-aware(vlan_filtering 1) bridge, and then create either a trunk or access port.
What you expected to happen:
Bridge ports to reflect the desired vlan ids or range in
bridge -c vlan show devoutputHow to reproduce it (as minimally and precisely as possible):
Apply below desiredState using NNCP:
Anything else we need to know?:
Using nmstatectl container from daemonset running on OpenShift cluster and providing the above desired state through yaml works fine!
Using nmcli directly from node works fine as well.
Using NNCP resource does not reflect right vlan ids(it always shows
2-4094unless you invokenmcli con upon the bridge port.(check below)bridge -c vlan show dev dummy0 dummy0 1 PVID Egress Untagged 2-4094Environment:
NodeNetworkStateon affected nodes (usekubectl get nodenetworkstate <node_name> -o yaml):NodeNetworkConfigurationPolicy:kubectl get pods --all-namespaces -l app=kubernetes-nmstate -o jsonpath='{.items[0].spec.containers[0].image}'):nmcli --version)kubectl version):The text was updated successfully, but these errors were encountered: