[ Kube 56 ] Install Calico network in Kubernetes

In this video I will show you how to provision a Kubernetes cluster with Calico as the overlay network. In all my videos, I used Flannel. But if you want to try some advanced networking feature, you will have to go with Calico or WeaveNet. Flannel only supports basic networking although its very simple and easy to install.
Github Link:
github.com/jus...
Learn Kubernetes Playlist:
• Learn Kubernetes
Hope you found this video useful. Please share it with your friends and make sure to subscribe to my channel. For any questions/issues/feedback, please leave me a comment and I will get back to you at the earliest I can.
Thanks for watching this video.
If you wish to support me:
www.paypal.com...
#kubernetes #k8s #justmeopensource #justmekubernetes

[@ramakris]

Great stuff dude. Keep up the good work.

[@justmeandopensource]

Hi Rama, thanks for watching. Cheers.

[@davidgillard8682]

Thanks for your videos

[@justmeandopensource]

Hi David, thanks for watching.

[@janardhanm6402]

Hai this is Janardhan, I like your videos and make video on " Networking policies" and Why we use the Networking Policies ??

[@justmeandopensource]

Hi Janardhan, thanks for your interest in my videos. I will add it to my list. Cheers.

[@janardhanm6402]

@@justmeandopensource Tanks, I have facing one error in the last 3 days but it's not resolved. My error is: curl:(28) failed to connect localhost port 31250: connection time out. MY port 31250 is listing what happened ???

[@sudheshpn]

Can you do a video on ingress/egress network policy using calico and a video on psp

[@justmeandopensource]

Hi Sudhesh, thanks for watching. Pod security Policy is in my list for a long time. I didn't get to do it so far. Will look into it time permitting. Cheers.

[@cd-stephen]

u sir - are amazing

[@saurabhagarwal5692]

Thank you so much for one more nice video. I have one question can we update our networking from flannel to Calico for running K8s cluster ?

[@justmeandopensource]

Hi Saurabh, thanks for watching. Yes you can change the overlay network from Flannel to Calico, but that will involve downtime as you will have to re-initialize the cluster components. Its wise to spend some time researching the capabilities of each overlay network and choosing the right one in first place. It will be hard to do this when you have lots of workloads in your cluster. Cheers.

[@saurabhagarwal5692]

Thank you so much for your response...

[@ronaldocorrea8007]

excellent!!!

[@justmeandopensource]

Hi Ronaldo, thanks for watching.

[@nikhildogra7706]

Hi, Can you please upload a detail video on kubernetes networking? Thx in advance

[@justmeandopensource]

Hi Nikhil, thanks for watching. I have different aspects of networking as separate topics and is in my to do list. Will get it covered at some point as I already have scheduled videos for the next couple months. Cheers.

[@guents]

I wish my calico installation worked as smoothly... I have wasted 2 days now, my pods won't talk to each other

[@rashivyas10]

Hi Venkat, have you covered RBAC in any of your video?

[@justmeandopensource]

HI Rashi, thanks for watching this video. I have used RBAC in most of my videos but haven't done any videos specifically about RBAC. I have that in my list and will get that recorded. Cheers.

[@rashivyas10]

@@justmeandopensource thank you for the reply Venkat,waiting for the video. :)

[@justmeandopensource]

Cheers.

[@henryhhl1]

Thanks for videos, is there any plan for istio/let’s encrypt?

[@justmeandopensource]

Hi Henry, thanks for watching this video. It's not in my list. I have few more basic Istio videos in my list and then will add this as well. Cheers.

[@ninja2807]

Hi. Thanks again for your fantastic videos. I have a question before deploying the calico network. On my Kubernetes cluster, either the master and worker node has two interfaces. 1st interface is nat, the second interface is bridged as I want to have an IP address from my local network. However, my local network IP range is 192.168.0.0/24. You mentioned that the calico works if the --pod-network-cidr=192.168.0.0/16, in this case, I believe I might have some problems because the networks will overlap. how can I change calico to work with a custom cidr range instead of the default 192.168.0.0/16.

[@justmeandopensource]

Hi Ninja2807, Thanks for watching this video and thanks for your interest in my channel. The default network pool (CIDR) Calico uses is 192.168.0.0/16. This is what you pass during the kubeadm init command like below. kubeadm init --pod-network-cidr=192.168.0.0/16 There are couple of ways to use custom network cidr for your pod range by Calico. I provisioned the cluster in VirtualBox VMs using my vagrant script. github.com/justmeandopensource/kubernetes/tree/master/vagrant-provisioning You can see in the below script how I am initializing the kubernetes and installing Calico. raw.githubusercontent.com/justmeandopensource/kubernetes/master/vagrant-provisioning/bootstrap_kmaster_calico.sh Basically after kubeadm init, I am deploying Calico network from the below manifest. docs.projectcalico.org/v3.11/manifests/calico.yaml In this manifest, you can change the ip pool before deploying. Also change the cidr during the kubeadm init command. If you already deployed Calico with default network, you can use calicoctl to change the pool. docs.projectcalico.org/v3.6/networking/changing-ip-pools Hope this helps.

[@ninja2807]

@@justmeandopensource thanks very much Venkat. I'll try this today.

[@justmeandopensource]

@@ninja2807 You are welcome.

[@expirytrader5802]

Hi it is been pure practical learning your channel. It been fantastic to here on your channel. I have condition where I wanted to connect multiple pods to single pods from multiple worker nodes AWS eks. How I can do that. Currently works good single worker node. Basically I'm connecting selenium nodes (pods) to selenium Hub (single pod). Which tool I will have to use ?

[@justmeandopensource]

Hi, thanks for watching. You should already be able to access every other pod if they are listening on any ports. I am not sure what do you mean by connecting multiple pods. But you should be able to do it as it is within the cluster. This is possible as we have overlay network that provides communication between pods that are running in different worker nodes.

[@expirytrader5802]

@@justmeandopensource ohh got resolved that AWS vpc has the cni plugin handles the enough ip. Yeah overlay network will be good choice in case with the lot of pods to interact with it.

[@justmeandopensource]

@@expirytrader5802 Cool.

[@hrishekessaraghueeti1258]

Could you please do a calico ipv6 ipv4 dual stack with kubernetes cluster

[@johandegraeve8187]

Thx a lot for the instructive video. I was wandering if you could help me out with a problem. I setup on my laptop 3 vm's with centos 7 and install kubernetes cluster on it. I used calico for the overlay network. No problems with installation but the liveness/readiness probes for the calico nodes on my worker nodes fail (felix is not ready and felix is not live) and cause these pods to restart. The log of the calico node on the worker node shows "Hit error connecting to datastore - retry error=Get 10.96.0.1:443/api/v1/nodes/foo: dial tcp 10.96.0.1:443: i/o timeout". Can you give me a clue of where to look?

[@justmeandopensource]

Hi Johan, thanks for watching. I have never encountered this problem in my environment. Do you have multiple interfaces in your 3 VMs? Are they virtualbox VMs? I have a vagrant environment that provisions a kubernetes cluster on CentOS 7 VMs and bootstraps k8s cluster. If you are interested you can give that a try. kzbin.info/www/bejne/rYHHenWbjK99qck github.com/justmeandopensource/kubernetes/tree/master/vagrant-provisioning

[@venkatk8314]

Hi venkat - can we have multiple calico nodes per worker node.?

[@justmeandopensource]

Hi Venkat, thanks for watching. Calico is deployed as a daemonset that runs on each node. You don't want to change the behaviour of this. What you mean by calido nodes and why do you want to run multiple of them in one node?

[@venkatk8314]

@@justmeandopensource say for example if one calico pod doesn't come up due to some reason, it is not completely started, in that case something like high availbility of calico i need.

[@justmeandopensource]

@@venkatk8314 You only need one calico pod per node and Kubernetes will make sure its running. If there is a problem with a calico pod in a node, then basically networking is broken in that node and pods on that node won't be able to communicate with other pods on another node. Which needs fixing.

[@venkatk8314]

@@justmeandopensource thank you

[@justmeandopensource]

@@venkatk8314 No worries.

[@ivandinata]

hi, i'm trying to implement a host with multiple NIC in it with calico, but always fail. can you give example on it. thanks

[@justmeandopensource]

Hi Ivan, thanks for watching. If you have multiple network interfaces on your machine, Calico will always pick the first one. If you have eth0, eth1, eth2 interfaces, it will pick eth0. You may have a different interface that routes to the internet eg, eth2. In which case, you should configure Calico to use eth2. All details for your use case is well explained in the below documentation. docs.projectcalico.org/networking/ip-autodetection

[@gnadha123]

Hi Venkat, have u shaRed any video abt k8s pods as jenkins slaves

[@justmeandopensource]

Hi, thanks for watching this video. I have done couple of videos on running Jenkins in Kubernetes. In those videos, I have shown how to run Jenkins slaves as kubernetes pods. I also have another video waiting to be released which is about connecting your existing external jenkins server to run slave pods in a kubernetes cluster. Please check the below links kzbin.info/www/bejne/hZOqg2N7m7uDosk kzbin.info/www/bejne/anabYZp7mtNgpZI Thanks

[@gnadha123]

@@justmeandopensource Thanks Venkat

[@justmeandopensource]

You are welcome. Cheers.

[@cactusfamily6441]

Hi I have quick question. Have you tried network policy ? I was trying to do it in Azure AKS cluster. It is not blocking with network policy it is suppose to.

[@justmeandopensource]

Thanks for watching. Haven't explore network policies. But thats in my list and will get it done sometime later.

[@cactusfamily6441]

Ok Thanks waiting your video about network policy :)

[@justmeandopensource]

@@cactusfamily6441 Cheers.

[@davehouser1]

I just cloned your repo, "boostrap_kmaster_calico.sh" does not appear to be listed anymore, was this removed?

[@justmeandopensource]

Hi Dave, thanks for watching. I used to use Flannel as the overlay network but switched to using Calico recently as it is more feature rich than Flannel. If you use my vagrant environment, it will by default use Calico now. The code this video is based on has been moved to a different branch. If you still want to have a look, here it is github.com/justmeandopensource/kubernetes/tree/2020 Cheers.

[@stilianstoilov3728]

Hi Venkat, what is the idea of running "\vi"?

[@justmeandopensource]

Hi Stilian, thanks for watching. I have vi aliased to vim in my zshrc file. Vim shows syntax highlighting. I wanted to turn off that temporarily so that viewers can clearly see the white text. I wanted to use vi instead of vim. \vi runs the actual vi without using the aliased version. $ alias vi=vim now vi will launch vim $ \vi now plain vi program

[@stilianstoilov3728]

@@justmeandopensource understood, 10x for the fast reply

[@justmeandopensource]

@@stilianstoilov3728 you are welcome.

[@avinashavinash9600]

I have a issue here when I user calico network its does not ping 2 different pods when its in different cluster (but it works when I check both pods are in same cluster) do you thing something is blocking here ? if yes please help me to know that

[@justmeandopensource]

Hi Avinash, thanks for watching. How is/are your cluster(s) setup? What do you mean by different cluster? You won't be able to talk to another pod on a different cluster anyway.

[@sameerpansari1912]

Hi Venkat , This is just beginning for me, amazing video presentation. But am using 2 ec2 instances (master and worker) , while executing any kubectl commands , error "The connection to the server localhost:8080 was refused - did you specify the right host or port? " might be silly , but very important to understood to move ahead.

[@kasimshaik]

Hi Venkat , i have been using weave as CNI plugin in many environments. Do you see any major differences between weave and calico. I prefer weave since we can define pod cidr range after bootstrapping first master not on kubeadm cmd line.

[@justmeandopensource]

Hi Kasim, Thanks for watching this video. I agree that Weave is more feature rich than Calico. I haven't explored much into the CNI networking side of Kubernetes. I had been using Flannel for a long time and then switched to Calico so that I can use pod network policies. As you mentioned, we have to define the pod cidr in kubeadm init command based on what network we use. I haven't tried weave net yet. There is an interesting blog post at rancher.com comparing different CNI providers. I can see weave net has rich features. I will try that in the future. rancher.com/blog/2019/2019-03-21-comparing-kubernetes-cni-providers-flannel-calico-canal-and-weave/ Thanks.

[@abhijitpaul6928]

What terminal is being used?

[@justmeandopensource]

Terminal emulator is Termite and shell is Zsh with few add ons on host Arch Linux. I did a video on my terminal setup if you are interested here it is kzbin.info/www/bejne/hoa6n3aYp56WhJo Thanks.

[@abhijitpaul6928]

@@justmeandopensource does it come for Mac...I liked in the video how you splited the borderless multiple terminal without switching tab.

[@justmeandopensource]

@@abhijitpaul6928 I think there are few ports of Termite for Mac. But you don't have to go with Termite. This can be achieved in any terminal.

[@thannasip8001]

we are going with weave net,is it better than weave net?

[@justmeandopensource]

Hi, I haven't actually tried Weave net yet. And you can't really say which one is better. You will have to try both and use whichever suits your needs. They are designed and developed for specific purposes. Cheers.