Ce contenu n'est pas disponible dans la langue sélectionnée.
Chapter 9. OVN-Kubernetes network plugin
9.1. About the OVN-Kubernetes network plugin
The Red Hat OpenShift Service on AWS cluster uses a virtualized network for pod and service networks.
Part of Red Hat OpenShift Networking, the OVN-Kubernetes network plugin is the default network provider for Red Hat OpenShift Service on AWS. OVN-Kubernetes is based on Open Virtual Network (OVN) and provides an overlay-based networking implementation. A cluster that uses the OVN-Kubernetes plugin also runs Open vSwitch (OVS) on each node. OVN configures OVS on each node to implement the declared network configuration.
OVN-Kubernetes is the default networking solution for Red Hat OpenShift Service on AWS and single-node OpenShift deployments.
OVN-Kubernetes, which arose from the OVS project, uses many of the same constructs, such as open flow rules, to determine how packets travel through the network. For more information, see the Open Virtual Network website.
OVN-Kubernetes is a series of daemons for OVS that translate virtual network configurations into OpenFlow
rules. OpenFlow
is a protocol for communicating with network switches and routers, providing a means for remotely controlling the flow of network traffic on a network device so that network administrators can configure, manage, and monitor the flow of network traffic.
OVN-Kubernetes provides more of the advanced functionality not available with OpenFlow
. OVN supports distributed virtual routing, distributed logical switches, access control, Dynamic Host Configuration Protocol (DHCP), and DNS. OVN implements distributed virtual routing within logic flows that equate to open flows. For example, if you have a pod that sends out a DHCP request to the DHCP server on the network, a logic flow rule in the request helps the OVN-Kubernetes handle the packet so that the server can respond with gateway, DNS server, IP address, and other information.
OVN-Kubernetes runs a daemon on each node. There are daemon sets for the databases and for the OVN controller that run on every node. The OVN controller programs the Open vSwitch daemon on the nodes to support the network provider features: egress IPs, firewalls, routers, hybrid networking, IPSEC encryption, IPv6, network policy, network policy logs, hardware offloading, and multicast.
9.1.1. OVN-Kubernetes purpose
The OVN-Kubernetes network plugin is an open-source, fully-featured Kubernetes CNI plugin that uses Open Virtual Network (OVN) to manage network traffic flows. OVN is a community developed, vendor-agnostic network virtualization solution. The OVN-Kubernetes network plugin uses the following technologies:
- OVN to manage network traffic flows.
- Kubernetes network policy support and logs, including ingress and egress rules.
- The Generic Network Virtualization Encapsulation (Geneve) protocol, rather than Virtual Extensible LAN (VXLAN), to create an overlay network between nodes.
The OVN-Kubernetes network plugin supports the following capabilities:
- Hybrid clusters that can run both Linux and Microsoft Windows workloads. This environment is known as hybrid networking.
- Offloading of network data processing from the host central processing unit (CPU) to compatible network cards and data processing units (DPUs). This is known as hardware offloading.
- IPv4-primary dual-stack networking on bare-metal, VMware vSphere, IBM Power®, IBM Z®, and RHOSP platforms.
- IPv6 single-stack networking on a bare-metal platform.
- IPv6-primary dual-stack networking for a cluster running on a bare-metal, a VMware vSphere, or an RHOSP platform.
- Egress firewall devices and egress IP addresses.
- Egress router devices that operate in redirect mode.
- IPsec encryption of intracluster communications.
9.1.2. OVN-Kubernetes IPv6 and dual-stack limitations
The OVN-Kubernetes network plugin has the following limitations:
For clusters configured for dual-stack networking, both IPv4 and IPv6 traffic must use the same network interface as the default gateway. If this requirement is not met, pods on the host in the
ovnkube-node
daemon set enter theCrashLoopBackOff
state. If you display a pod with a command such asoc get pod -n openshift-ovn-kubernetes -l app=ovnkube-node -o yaml
, thestatus
field contains more than one message about the default gateway, as shown in the following output:I1006 16:09:50.985852 60651 helper_linux.go:73] Found default gateway interface br-ex 192.168.127.1 I1006 16:09:50.985923 60651 helper_linux.go:73] Found default gateway interface ens4 fe80::5054:ff:febe:bcd4 F1006 16:09:50.985939 60651 ovnkube.go:130] multiple gateway interfaces detected: br-ex ens4
The only resolution is to reconfigure the host networking so that both IP families use the same network interface for the default gateway.
For clusters configured for dual-stack networking, both the IPv4 and IPv6 routing tables must contain the default gateway. If this requirement is not met, pods on the host in the
ovnkube-node
daemon set enter theCrashLoopBackOff
state. If you display a pod with a command such asoc get pod -n openshift-ovn-kubernetes -l app=ovnkube-node -o yaml
, thestatus
field contains more than one message about the default gateway, as shown in the following output:I0512 19:07:17.589083 108432 helper_linux.go:74] Found default gateway interface br-ex 192.168.123.1 F0512 19:07:17.589141 108432 ovnkube.go:133] failed to get default gateway interface
The only resolution is to reconfigure the host networking so that both IP families contain the default gateway.
9.1.3. Session affinity
Session affinity is a feature that applies to Kubernetes Service
objects. You can use session affinity if you want to ensure that each time you connect to a <service_VIP>:<Port>, the traffic is always load balanced to the same back end. For more information, including how to set session affinity based on a client’s IP address, see Session affinity.
Stickiness timeout for session affinity
The OVN-Kubernetes network plugin for Red Hat OpenShift Service on AWS calculates the stickiness timeout for a session from a client based on the last packet. For example, if you run a curl
command 10 times, the sticky session timer starts from the tenth packet not the first. As a result, if the client is continuously contacting the service, then the session never times out. The timeout starts when the service has not received a packet for the amount of time set by the timeoutSeconds
parameter.
9.2. Configuring an egress IP address
As a cluster administrator, you can configure the OVN-Kubernetes Container Network Interface (CNI) network plugin to assign one or more egress IP addresses to a namespace, or to specific pods in a namespace.
9.2.1. Egress IP address architectural design and implementation
The Red Hat OpenShift Service on AWS egress IP address functionality allows you to ensure that the traffic from one or more pods in one or more namespaces has a consistent source IP address for services outside the cluster network.
For example, you might have a pod that periodically queries a database that is hosted on a server outside of your cluster. To enforce access requirements for the server, a packet filtering device is configured to allow traffic only from specific IP addresses. To ensure that you can reliably allow access to the server from only that specific pod, you can configure a specific egress IP address for the pod that makes the requests to the server.
An egress IP address assigned to a namespace is different from an egress router, which is used to send traffic to specific destinations.
In ROSA with HCP clusters, application pods and ingress router pods run on the same node. If you configure an egress IP address for an application project in this scenario, the IP address is not used when you send a request to a route from the application project.
The assignment of egress IP addresses to control plane nodes with the EgressIP feature is not supported.
The following examples illustrate the annotation from nodes on several public cloud providers. The annotations are indented for readability.
Example cloud.network.openshift.io/egress-ipconfig
annotation on AWS
cloud.network.openshift.io/egress-ipconfig: [ { "interface":"eni-078d267045138e436", "ifaddr":{"ipv4":"10.0.128.0/18"}, "capacity":{"ipv4":14,"ipv6":15} } ]
The following sections describe the IP address capacity for supported public cloud environments for use in your capacity calculation.
9.2.1.1. Amazon Web Services (AWS) IP address capacity limits
On AWS, constraints on IP address assignments depend on the instance type configured. For more information, see IP addresses per network interface per instance type
9.2.1.2. Assignment of egress IPs to pods
To assign one or more egress IPs to a namespace or specific pods in a namespace, the following conditions must be satisfied:
-
At least one node in your cluster must have the
k8s.ovn.org/egress-assignable: ""
label. -
An
EgressIP
object exists that defines one or more egress IP addresses to use as the source IP address for traffic leaving the cluster from pods in a namespace.
If you create EgressIP
objects prior to labeling any nodes in your cluster for egress IP assignment, Red Hat OpenShift Service on AWS might assign every egress IP address to the first node with the k8s.ovn.org/egress-assignable: ""
label.
To ensure that egress IP addresses are widely distributed across nodes in the cluster, always apply the label to the nodes you intent to host the egress IP addresses before creating any EgressIP
objects.
9.2.1.3. Assignment of egress IPs to nodes
When creating an EgressIP
object, the following conditions apply to nodes that are labeled with the k8s.ovn.org/egress-assignable: ""
label:
- An egress IP address is never assigned to more than one node at a time.
- An egress IP address is equally balanced between available nodes that can host the egress IP address.
If the
spec.EgressIPs
array in anEgressIP
object specifies more than one IP address, the following conditions apply:- No node will ever host more than one of the specified IP addresses.
- Traffic is balanced roughly equally between the specified IP addresses for a given namespace.
- If a node becomes unavailable, any egress IP addresses assigned to it are automatically reassigned, subject to the previously described conditions.
When a pod matches the selector for multiple EgressIP
objects, there is no guarantee which of the egress IP addresses that are specified in the EgressIP
objects is assigned as the egress IP address for the pod.
Additionally, if an EgressIP
object specifies multiple egress IP addresses, there is no guarantee which of the egress IP addresses might be used. For example, if a pod matches a selector for an EgressIP
object with two egress IP addresses, 10.10.20.1
and 10.10.20.2
, either might be used for each TCP connection or UDP conversation.
9.2.1.4. Architectural diagram of an egress IP address configuration
The following diagram depicts an egress IP address configuration. The diagram describes four pods in two different namespaces running on three nodes in a cluster. The nodes are assigned IP addresses from the 192.168.126.0/18
CIDR block on the host network.
Both Node 1 and Node 3 are labeled with k8s.ovn.org/egress-assignable: ""
and thus available for the assignment of egress IP addresses.
The dashed lines in the diagram depict the traffic flow from pod1, pod2, and pod3 traveling through the pod network to egress the cluster from Node 1 and Node 3. When an external service receives traffic from any of the pods selected by the example EgressIP
object, the source IP address is either 192.168.126.10
or 192.168.126.102
. The traffic is balanced roughly equally between these two nodes.
The following resources from the diagram are illustrated in detail:
Namespace
objectsThe namespaces are defined in the following manifest:
Namespace objects
apiVersion: v1 kind: Namespace metadata: name: namespace1 labels: env: prod --- apiVersion: v1 kind: Namespace metadata: name: namespace2 labels: env: prod
EgressIP
objectThe following
EgressIP
object describes a configuration that selects all pods in any namespace with theenv
label set toprod
. The egress IP addresses for the selected pods are192.168.126.10
and192.168.126.102
.EgressIP
objectapiVersion: k8s.ovn.org/v1 kind: EgressIP metadata: name: egressips-prod spec: egressIPs: - 192.168.126.10 - 192.168.126.102 namespaceSelector: matchLabels: env: prod status: items: - node: node1 egressIP: 192.168.126.10 - node: node3 egressIP: 192.168.126.102
For the configuration in the previous example, Red Hat OpenShift Service on AWS assigns both egress IP addresses to the available nodes. The
status
field reflects whether and where the egress IP addresses are assigned.
9.2.2. EgressIP object
The following YAML describes the API for the EgressIP
object. The scope of the object is cluster-wide; it is not created in a namespace.
apiVersion: k8s.ovn.org/v1 kind: EgressIP metadata: name: <name> 1 spec: egressIPs: 2 - <ip_address> namespaceSelector: 3 ... podSelector: 4 ...
- 1
- The name for the
EgressIPs
object. - 2
- An array of one or more IP addresses.
- 3
- One or more selectors for the namespaces to associate the egress IP addresses with.
- 4
- Optional: One or more selectors for pods in the specified namespaces to associate egress IP addresses with. Applying these selectors allows for the selection of a subset of pods within a namespace.
The following YAML describes the stanza for the namespace selector:
Namespace selector stanza
namespaceSelector: 1
matchLabels:
<label_name>: <label_value>
- 1
- One or more matching rules for namespaces. If more than one match rule is provided, all matching namespaces are selected.
The following YAML describes the optional stanza for the pod selector:
Pod selector stanza
podSelector: 1
matchLabels:
<label_name>: <label_value>
- 1
- Optional: One or more matching rules for pods in the namespaces that match the specified
namespaceSelector
rules. If specified, only pods that match are selected. Others pods in the namespace are not selected.
In the following example, the EgressIP
object associates the 192.168.126.11
and 192.168.126.102
egress IP addresses with pods that have the app
label set to web
and are in the namespaces that have the env
label set to prod
:
Example EgressIP
object
apiVersion: k8s.ovn.org/v1 kind: EgressIP metadata: name: egress-group1 spec: egressIPs: - 192.168.126.11 - 192.168.126.102 podSelector: matchLabels: app: web namespaceSelector: matchLabels: env: prod
In the following example, the EgressIP
object associates the 192.168.127.30
and 192.168.127.40
egress IP addresses with any pods that do not have the environment
label set to development
:
Example EgressIP
object
apiVersion: k8s.ovn.org/v1 kind: EgressIP metadata: name: egress-group2 spec: egressIPs: - 192.168.127.30 - 192.168.127.40 namespaceSelector: matchExpressions: - key: environment operator: NotIn values: - development
9.2.3. Labeling a node to host egress IP addresses
You can apply the k8s.ovn.org/egress-assignable=""
label to a node in your cluster so that Red Hat OpenShift Service on AWS can assign one or more egress IP addresses to the node.
Prerequisites
-
Install the ROSA CLI (
rosa
). - Log in to the cluster as a cluster administrator.
Procedure
To label a node so that it can host one or more egress IP addresses, enter the following command:
$ rosa edit machinepool <machinepool_name> --cluster=<cluster_name> --labels "k8s.ovn.org/egress-assignable="
ImportantThis command replaces any exciting node labels on your machinepool. You should include any of the desired labels to the
--labels
field to ensure that your existing node labels persist.