Install-CNI-plugin

Verifying Network IP Ranges

Checking for IP Range Overlap

[!IMPORTANT] The pod network CIDR block must not overlap with the node IP address range.

Node IP Range

[!INFO] Nodes get IP addresses from the VPC (private network) in AWS:

• Check VPC service to see IP address range
• Example: 172.16.x.x (may vary by network)

Pod Network Range

[!INFO] Weavenet default configuration:

• Default range: 10.32.0.0/12
• Capacity: ~1 million IP addresses
• Range is evenly divided between cluster nodes

Validation

[!NOTE] In this case, the pod network CIDR (10.32.0.0/12) and node CIDR do not overlap, so we can use Weavenet’s default IP range.

[!TIP] To verify CIDR ranges:

• Use any CIDR block calculator
• Input the CIDR block
• Review full IP address range
• Confirm no overlap between pod and node ranges

Installing Weavenet CNI Plugin

Finding the Installation Guide

[!INFO] Weavenet can be found in the official Kubernetes documentation under network add-ons, along with other CNI plugin options.

Installation Methods

Quick Install

[!NOTE] Weavenet provides a one-line installation command that:

• Uses kubectl apply to deploy the plugin
• Applies a Kubernetes manifest containing all required components

Manual Installation

[!TIP] For better version control and configuration:

1. Download the manifest file locally:

   wget <weavenet-url> -O weave.yaml

2. Review the manifest contents before applying

Understanding the Manifest

[!INFO] The manifest creates several Kubernetes components:

• Main component is a DaemonSet
• Weavenet application runs on port 6784

Configuring the CIDR Block

Default Configuration

[!NOTE] Weavenet uses 10.32.0.0/12 as the default CIDR block

Custom Configuration

[!TIP] To override the default CIDR block:

1. Locate the DaemonSet component in the YAML
2. Find the main weave-kube container
3. Modify the launch command parameters

[!IMPORTANT] When modifying the YAML:

• Commands are specified as a list in YAML format
• Each command option should be on a new line
• Example format:

  command:
    - /home/weave/launch.sh
    - --ipalloc-range=10.32.1.0/12
    - <additional-options>

And now we can apply which

will install all these components including the daemon set.

First, our master node should be in a ready state,

so let’s check that kubectl get node.

And as you see, we have status ready for master,

so that one’s fixed. And the second

one was that core DNS pods were not starting,

they were in a pending state. So that should be fixed as

well. And let’s see that get POD in

cube system namespace.

And there you go. We have coredns pods both

in a running status as well as we have a new pod

here we’ve net on our masternode. And since

we just have one node in the cluster, we have one weave

pod that basically manages the POD network in our

cluster.

And finally we’ve talked about POD network

and POD IPs. So how do we even see

the IP addresses of the pods in the cluster. Well, one way of

doing it is by getting detailed output

of the pod. So let’s check the

corednspod and describe

that pod in detail with kubectl,

describe POD command and the POD name as well

as the namespace.

And this will give us a bunch of output of containers

creating and starting inside the pod,

as well as some metadata of the pod. And as you see,

we have an IP address for corednspod which

is in the range of the CIDR block

that we defined for weavenet, which is132.03.

So that’s one way to check the IP address of a pod.

However, doing this for every pod in the cluster

is too much work. And also you don’t have a nice overview

of list of pods with their IP addresses. To do that,

we can display the pods using output

option dash O, which stands for output option

wide. So this will basically give us an extended

output of the pods. So let’s execute.

And there you go. We have additional columns with additional

information and one of them is the IP address.

And as you see, the CoreDNS pods,

which started after we’ve got deployed,

have the IP addresses both from the range of

WaveNet’s pod networks CIDR block. And you also

notice that all the other pods, the static pods

as well as Kube proxy and the weavenet itself,

have a different IP address which is the IP

address of the control plane node they are

running on. And that’s because these pods are not

part of the regular pod network like CoreDNS

for example, or any other POD application pods,

database pods that we’re going to deploy later.

And that’s why they get these static IP

address of the node. And to check that we can

also do kubectl get node with wide output

and see the internal IP address of the master node,

which is same as right here. So now we have our

cluster completely set up, we have our control plane processes

running and we have a POD network deployed in

the cluster. However, right now we just have a one node cluster

with only a master node. So it’s time to join the worker

nodes and make them part of the cluster. And installing the

POD network plugin was needed to

add the worker nodes to the cluster, otherwise we wouldn’t be able to join

them and create pod networks on them. So now

that we have everything prepared, let’s configure and join

the worker nodes.