Enable network policy support functionality¶
Enable policy, pod_label and namespace handlers to respond to network policy events. As this is not done by default you’d have to explicitly add that to the list of enabled handlers at kuryr.conf (further info on how to do this can be found at Containerized installation):
[kubernetes]
enabled_handlers=vif,lb,lbaasspec,policy,pod_label,namespace,kuryrnetwork,kuryrnetworkpolicy,kuryrport
Note that if you also want to enable prepopulation of ports pools upon new namespace creation, you need to also dd the kuryrnetwork_population handler (more details on How to enable ports pool support):
[kubernetes]
enabled_handlers=vif,lb,lbaasspec,policy,pod_label,namespace,kuryrnetworkpolicy,kuryrnetwork,kuryrnetwork_population,kuryrport
After that, enable also the security group drivers for policies:
[kubernetes]
service_security_groups_driver = policy
pod_security_groups_driver = policy
Warning
The correct behavior for pods that have no network policy applied is to
allow all ingress and egress traffic. If you want that to be enforced,
please make sure to create an SG allowing all traffic and add it to
[neutron_defaults]pod_security_groups
setting in kuryr.conf
:
[neutron_defaults]
pod_security_groups = ALLOW_ALL_SG_ID
Enable the namespace subnet driver by modifying the default pod_subnet_driver option:
[kubernetes]
pod_subnets_driver = namespace
Select the subnet pool from where the new subnets will get their CIDR:
[namespace_subnet]
pod_subnet_pool = SUBNET_POOL_ID
Lastly, select the router where the new subnet will be connected:
[namespace_subnet]
pod_router = ROUTER_ID
Note you need to restart the kuryr controller after applying the above step. For devstack non-containerized deployments:
$ sudo systemctl restart devstack@kuryr-kubernetes.service
Same for containerized deployments:
$ kubectl -n kube-system get pod | grep kuryr-controller
$ kubectl -n kube-system delete pod KURYR_CONTROLLER_POD_NAME
For directly enabling the driver when deploying with devstack, you just need to add the policy, pod_label and namespace handler and drivers with:
KURYR_ENABLED_HANDLERS=vif,lb,lbaasspec,policy,pod_label,namespace,kuryrnetworkpolicy,kuryrport
KURYR_SG_DRIVER=policy
KURYR_SUBNET_DRIVER=namespace
Note
If the loadbalancer maintains the source IP (such as ovn-octavia driver), there is no need to enforce sg rules at the load balancer level. To disable the enforcement, you need to set the following variable:
KURYR_ENFORCE_SG_RULES=False
Testing the network policy support functionality¶
Given a yaml file with a network policy, such as:
apiVersion: networking.k8s.io/v1 kind: NetworkPolicy metadata: name: test-network-policy namespace: default spec: podSelector: matchLabels: project: default policyTypes: - Ingress - Egress ingress: - from: - namespaceSelector: matchLabels: project: default ports: - protocol: TCP port: 6379 egress: - to: - namespaceSelector: matchLabels: project: default ports: - protocol: TCP port: 5978
Apply the network policy:
$ kubectl apply -f network_policy.yml
Check that the resources has been created:
$ kubectl get kuryrnetworkpolicies NAME AGE test-network-policy 2s $ kubectl get networkpolicies NAME POD-SELECTOR AGE test-network-policy role=db 2s $ openstack security group list | grep sg-test-network-policy | dabdf308-7eed-43ef-a058-af84d1954acb | sg-test-network-policy
Check that the rules are in place for the security group:
$ kubectl get kuryrnetworkpolicy test-network-policy -o yaml apiVersion: openstack.org/v1 kind: KuryrNetworkPolicy metadata: annotations: networkPolicyLink: <link> clusterName: "" creationTimestamp: 2018-10-02T11:17:02Z generation: 0 name: test-network-policy namespace: default resourceVersion: "2117" selfLink: /apis/openstack.org/v1/namespaces/default/kuryrnetworkpolicies/test-network-policy uid: afb99326-c634-11e8-b63d-002564fdd760 spec: egressSgRules: - sgRule: description: Kuryr-Kubernetes NetPolicy SG rule direction: egress ethertype: IPv4 port_range_max: 5978 port_range_min: 5978 protocol: tcp ingressSgRules: - sgRule: description: Kuryr-Kubernetes NetPolicy SG rule direction: ingress ethertype: IPv4 port_range_max: 6379 port_range_min: 6379 protocol: tcp status: securityGroupId: cdee7815-3b49-4a3e-abc8-31e384ab75c5 securityGroupRules: … $ openstack security group rule list sg-test-network-policy --protocol tcp -c "IP Protocol" -c "Port Range" -c "Direction" --long +-------------+------------+-----------+ | IP Protocol | Port Range | Direction | +-------------+------------+-----------+ | tcp | 6379:6379 | ingress | | tcp | 5978:5978 | egress | +-------------+------------+-----------+
Create a pod:
$ kubectl create deployment --image quay.io/kuryr/demo demo deployment "demo" created $ kubectl get pod -o wide NAME READY STATUS RESTARTS AGE IP demo-5558c7865d-fdkdv 1/1 Running 0 44s 10.0.0.68
Get the pod port and check its security group rules:
$ openstack port list --fixed-ip ip-address=10.0.0.68 -f value -c ID 5d29b83c-714c-4579-8987-d0c0558420b3 $ openstack port show 5d29b83c-714c-4579-8987-d0c0558420b3 | grep security_group_ids | security_group_ids | bb2ac605-56ff-4688-b4f1-1d045ad251d0 $ openstack security group rule list bb2ac605-56ff-4688-b4f1-1d045ad251d0 --protocol tcp -c "IP Protocol" -c "Port Range" +-------------+------------+-----------+ | IP Protocol | Port Range | Direction | +-------------+------------+-----------+ | tcp | 6379:6379 | ingress | | tcp | 5978:5978 | egress | +-------------+------------+-----------+
Try to curl the pod on port 8080 (hint: it won’t work!):
$ curl 10.0.0.68:8080
Update network policy to allow ingress 8080 port:
$ kubectl patch networkpolicy test-network-policy -p '{"spec":{"ingress":[{"ports":[{"port": 8080,"protocol": "TCP"}]}]}}' networkpolicy "test-network-policy" patched $ kubectl get knp test-network-policy -o yaml apiVersion: openstack.org/v1 kind: KuryrNetworkPolicy metadata: annotations: networkPolicyLink: <link> clusterName: "" creationTimestamp: 2018-10-02T11:17:02Z generation: 0 name: test-network-policy namespace: default resourceVersion: "1546" selfLink: /apis/openstack.org/v1/namespaces/default/kuryrnetworkpolicies/np-test-network-policy uid: afb99326-c634-11e8-b63d-002564fdd760 spec: egressSgRules: - sgRule: description: Kuryr-Kubernetes NetPolicy SG rule direction: egress ethertype: IPv4 port_range_max: 5978 port_range_min: 5978 protocol: tcp ingressSgRules: - sgRule: description: Kuryr-Kubernetes NetPolicy SG rule direction: ingress ethertype: IPv4 port_range_max: 8080 port_range_min: 8080 protocol: tcp status: securityGroupId: cdee7815-3b49-4a3e-abc8-31e384ab75c5 securityGroupRules: … $ openstack security group rule list sg-test-network-policy -c "IP Protocol" -c "Port Range" -c "Direction" --long +-------------+------------+-----------+ | IP Protocol | Port Range | Direction | +-------------+------------+-----------+ | tcp | 8080:8080 | ingress | | tcp | 5978:5978 | egress | +-------------+------------+-----------+
Try to curl the pod ip after patching the network policy:
$ curl 10.0.0.68:8080 demo-5558c7865d-fdkdv: HELLO! I AM ALIVE!!!
Note the curl only works from pods (neutron ports) on a namespace that has the label project: default as stated on the policy namespaceSelector.
We can also create a single pod, without a label and check that there is no connectivity to it, as it does not match the network policy podSelector:
$ cat sample-pod.yml apiVersion: v1 kind: Pod metadata: name: demo-pod spec: containers: - image: quay.io/kuryr/demo imagePullPolicy: Always name: demo-pod $ kubectl apply -f sample-pod.yml $ curl demo-pod-IP:8080 NO REPLY
If we add to the pod a label that match a network policy podSelector, in this case ‘project: default’, the network policy will get applied on the pod, and the traffic will be allowed:
$ kubectl label pod demo-pod project=default $ curl demo-pod-IP:8080 demo-pod-XXX: HELLO! I AM ALIVE!!!
Confirm the teardown of the resources once the network policy is removed:
$ kubectl delete -f network_policy.yml $ kubectl get kuryrnetworkpolicies $ kubectl get networkpolicies $ openstack security group list | grep sg-test-network-policy