Networking

Routing, Switching, Gateways

To find gateway:

route
# or
ip route list

To add entries into the routing table. Where 2nd ip is gateway

ip route add 192.168.1.0/24 via 192.168.2.1

If forwarding between machines required for communication without a router:

/proc/sys/net/ipv4/ip_forward must be enabled:

echo 1 > /proc/sys/net/ipv4/ip_forward

To persist:

# /etc/sysctl.conf
net.ipv4.ip_forward=1

DNS

In /etc/resolv.conf set DNS server:

nameserver 192.168.1.100

We can set order for /etc/hosts or DNS server in:

/etc/nsswitch.conf:

hosts:  files dns

We can use nslookup, dig to query DNS servers:

nslookup google.ca
dig google.ca

Network Namespaces

Lets us have isolated routing and arp tables along with virtual interfaces.

ip netns add b
ip netns list

Run in ns:

ip netns exec red ip link
# or
ip -n res link

To connect namespaces we can use a virtual pair (or pipe):

To create a virtual cable

ip link add veth-red type veth peer name veth-blue

To attach with the network namespaces

ip link set veth-red netns red
ip link set veth-blue netns blue

To add an IP address

ip -n red addr add 192.168.15.1/24 dev veth-red
ip -n blue addr add 192.168.15.2/24 dev veth-blue

To set up ns interfaces

ip -n red link set veth-red up
ip -n blue link set veth-blue up

Check the connectivity

ip netns exec red ping 192.168.15.2

When we have many NS, we create a switch (bridge)

Putting this all together we can have the bridge reach the external network by talking to our host as the Gateway, and have connections go back in to the private network by implementing NAT on our host via:

iptables -t nat -A PREROUTING --dport 80 --to-destination 192.168.15.2:80 -j DNAT

Pod Networking

The rules of kubernetes pod networking are that:

every pod should have an IP address
every pod should be able to communicate with every other pod in the same node
every pod should be able to communicate with every other pod on other nodes without NAT

We create a bridge on each node for the containers. Each bridge has a private subnet. To allow cross-node communications we add routes between nodes or use a router.

See kube-controller-manager --cluster-cidr= for pod range.

CNI in Kubernetes

We specify CNI plugin on container runtime in /etc/cni/net.d, bins in /opt/cni/bin

Kubernetes networking Solutions will typically install agents on every node (DaemonSet) along with bridges and then deal with peer-to-peer communication

IP Address Managements (IPAM)

Who assigns IPs to containers. CNI plugin manages the IP management.

Service Networking

Pods communicate via services and each gets a cluster-wide IP.

kube-proxy watches for service creation, and creates one. This is done by each node setting up forwarding rules on each node.

proxy-mode defines how forwarding rules are created on kube-proxy.

service-cluster-ip-range defines ip range for services.

ps -aux | grep kube-apiserver

--secure-port=6443 --service-account-key-file=/etc/kubernetes/pki/sa.pub --
service-cluster-ip-range=10.96.0.0/12

DNS in Kubernetes

Whenever we create services they get a DNS entry so any pod can access.

If in same namespace, can use just service name, eg 'service'

In different namespace add namespace suffix, eg 'service.default'.

Full domain is 'service.default.svc' or FQDN 'service.default.svc.cluster.local'

By default pods do not get entry, but we can enable DNS enties for them, thry get entry:

'IP-WITH-DASHES.namespace.pod'

eg '10-244-2-5.default.pod'

CoreDNS

Config in /etc/coredns/Corefile

.:53 {
    errors
    health {       lameduck 5s
    }
    ready
    kubernetes cluster.local in-addr.arpa ip6.arpa {
       pods insecure
       fallthrough in-addr.arpa ip6.arpa
       ttl 30
    }
    prometheus :9153
    forward . /etc/resolv.conf
    cache 30
    loop
    reload
}

kubectl get configmap -n kube-system

Kubelet configures DNS server for pods by setting nameserver in /etc/resolv.conf

resolv.conf also contains a search query to allow PARTIAL FQDN:

search default.svc.cluster.local svc.cluster.local cluster.local

Ingress Controllers

Native internal loadbalancing.

Not deployed by default.

GCE, nginx maintained by k8s (currently)

Create an ingress service account, and service.

Create a deployment with:

kind: ConfigMap
apiVersion: v1
metadata:
  name: nginx-configuration

apiVersion: apps/v1
kind: Deployment
metadata:
  name: ingress-controller
spec:
  replicas: 1
  selector:
    matchLabels:
      name: nginx-ingress
  template:
    metadata:
      labels:
        name: nginx-ingress
    spec:
      serviceAccountName: ingress-serviceaccount
      containers:
        - name: nginx-ingress-controller
          image: quay.io/kubernetes-ingress-controller/nginx-ingress-controller:0.21.0
          args:
            - /nginx-ingress-controller
            - --configmap=$(POD_NAMESPACE)/nginx-configuration
          env:
            - name: POD_NAME
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            - name: POD_NAMESPACE
              valueFrom:
                fieldRef:
                  fieldPath: metadata.namespace
          ports:
            - name: http
              containerPort: 80
            - name: https
              containerPort: 443

To configure ingress, create an ingress-resource:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: ingress-wear
spec:
     backend:
        serviceName: wear-service
        servicePort: 80

kubectl get ingress

To define rules for paths:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: ingress-wear-watch
spec:
  rules:
  - http:
      paths:
      - path: /wear
        backend:
          serviceName: wear-service
          servicePort: 80
      - path: /watch
        backend:
          serviceName: watch-service
          servicePort: 80

For domain name rules:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: ingress-wear-watch
spec:
  rules:
  - host: wear.my-online-store.com
    http:
      paths:
      - backend:
          serviceName: wear-service
          servicePort: 80
  - host: watch.my-online-store.com
    http:
      paths:
      - backend:
          serviceName: watch-service
          servicePort: 80

PreviousStorage NextInstall Kubernetes with kubeadm

Last updated 9 months ago