kubernetes-training

Networking Model

• Every pod should have an IP address
• Every pod should be able to talk to other pods within same node and other nodes as well (w/o NAT)

Setting up cluster : 1 master + 2 workers (manual)

• GIven node 1: .1.11, node 2: .1.12, node 3: .1.13
• Each node: ip link set v-net-0 type bridge && ip link set dev v-net-0 up
• Set the ip address for the bridge interface, each bridge is on its own subnet: ip addr add 10.224.1.1/24 dev v-net-0
• For each of the containers added into the node, run a script:
  • create veth pair
  • attach to ns
  • assign ip address, add route
  • bring up the interfaces

Inter-node connection

	$bluepod ping 10.244.2.2 # -> failed 
	$node1 ip route add 10.244.2.2 via 192.168.1.12 #(node2's ip)

• This gets complicated as more pods added.
• Solution: doing via the router, points host to the router
• All bridge’s subnet joined under 1 subnet: 10.244.1.0/24 -> 10.244.0.0/16

CNI

• Helps with all of the manual steps performed above
• CNI contains:
  • Add section: taking care of network associations with adding containers to the networks
  • Delete section: taking care of deleting network associations with deleting containers to the networks

    Container runtime

• Responsible for creating container
• /etc/cni/net.d/net-script.conflist : identify plugin
• /opt/cni/bin/net-script.sh: to find the script
• ./net-script.sh add <cntainer> <namespace> : script executed, adding container to the network namespace

[!Tip]

• --service-cluster-ip-range=172.20.0.0/16 flag configured in the kube-controller-manager and kube-api is service’s ip range
• --cluster-cidr in kubectl cluster-info dump is pod’s ip range