Dude. You're such a genius. Don't know how I spent the last 33 mins watching this video. I love the way you explain the concepts with such great detail. This is my first video of your channel. I am already a great fan of yours. Keep up the spirit and spread the knowledge for free.
@@justmeandopensource First time subbie here. Just discovered your content tonight and I'm wondering why TH-cam Algo hasn't done the needful since. You're so talented, Bro. Will be recommending and following up all your stuff.
@@justmeandopensource I was thinking on the upgrade process and it just so happens that I have to upgrade a cluster from 1.15.x to 1.16.x, and an important piece of that upgrade is the CNI part, in my case, maybe going from a somewhat ancient version of Calico to a newer one. Not saying that you cover exactly my use case, but if you can cover how to maintain things working WHILE also upgrading the CNI portion of the cluster, I guess that would be really really useful and it's a real situation on lots of production systems. Thanks, loving the content on your channel ...
I have given image id, then it start working. Very good video. Thank you. command given : kops create cluster --name justme.k8s.local --zones eu-west-1a --master-size t2.micro --node-size t2.micro --image "ami-0eb11ab33f229b26c" --kubernetes-version 1.15.0
Hi James, thanks for watching. Apologies for not clarifying about DNS in this video. I didn't register any domain instead used gossip-based domain. Basically anything ending with k8s.local. In this video I used justme.k8s.local kops.sigs.k8s.io/getting_started/aws/#prepare-local-environment
Kops doesn't create AWS EKS managed kubernetes cluster but just set of ec2 instances and kubernetes on top of it. So you won't see the cluster in EKS console. Just check the ec2 instances.
Hi Venkat, Great channel! Forgive my ignorance, but for the high availability (HA) masters set-up, how does KOPS make it possible to use a single ELB Classic load balancer to serve traffic to two different Auto Scaling Groups each with their own master instances? I didn't think this was possible unless an ALB was used. What am I missing in my understanding? Thanks!
Thanks for this video..I'am new to devops and want to switch job in devops position, but its getting very difficult for me to prepare for technical interviews since I dont have real time experience of devops culture and work. I would be very grateful to you if you would provide some overview of how to prepare for devops technical interview.What should be the preparation areas..topics to cover..not the technicalities or technical details but the overall idea of how should be the preparation points and topics. It would be a much help. Many many Thankyou in advance.
Hi Surendar, thanks for watching. In all my Kubernetes provisioning videos, I used either flannel or Calico as the overlay network. Installing Weave as CNI is very similar. Apart from installing the overlay network, I haven't looked much into the features of inidividual providers. Soon I will be exploring pod network policies which requires Calico or Weave as the overlay network. Cheers.
Hello Venkat, Thank you for posting this video. I was trying to implement kops following your step-by-step video but couldn't connect to the master and nodes via ssh. It throws a "permission denied(publickey) error. How do I fix this as I didn't specify the key while creating the cluster? is it possible to generate a key and apply it after the cluster has been already created?
Hi, could you please tell me when we create a cluster (master and nodes) using kops it's default Os is ubuntu or something else. so my question is how can we give a specific Linux Os for both master and nods excepts default os. I mean can we change the default os as a cento or Red Hat Linux ? .thanx
Hi KLP, thank for watching. By default, if you set up a cluster using KOPS as of today, the default OS will be Ubuntu 20.04. KOPS supports a lot of Linux distribution. Take a look at the below documentation on how to change the OS. You basically have to create your own AMI with one of the supported OS and the edit kops instance group to use a particular image. github.com/kubernetes/kops/blob/ed32f19591769a712abfb2fb52f37adc5f3aba44/docs/operations/images.md Cheers.
Hi, could you please explain what will happen if the master node got terminated? I had an experienced, my master node terminated and recreated it self help of an autoscaling group.but after recreated a new master node, my cluster is not worked perfectly like previous(FYI: I used AWS dynamic Ip address. after the recreated master node IP address also changed)
Hi KLP, thanks for watching. I haven't tested that scenario. The node instance group is responsible for maintaining the specified number of nodes in that group. So if one of your master died/crashed, the instance group will bring up another one but I believe you will have to run KOPS again to set up that new master node. I am not entirely sure about it.
Hi Venkat, Is it possible for u to explain dual IP stack support in kubernetes. Maybe some demo about how to enable the same and also validate if a pod has both IP addresses.
Hi Amartya, thanks for watching. I have read about the dual stack but have never explored it. I have videos for the next two months scheduled. Meanwhile when I get a chance I will explore this. Cheers.
Hi Venkat. thanks much . It’s magic for creating k8s cluster now using kops. I tried creating cluster with private topology and bastion. And the cluster is ready. However I am not sure how to do ssh to the master and worker nodes as they are not having public up. Pls assist. Thanks.
Hi, thanks for watching. I haven't tried that but did you use the option --ssh-public-key when creating the cluster using kops create cluster command? You may have already seen this article but if you haven't, medium.com/andcloudio/kubernetes-kops-cluster-on-aws-f55d197d8304
Hello Venkat, I followed the above steps on my newly created AWS account , I am getting the error "SSH public key must be specified when running with AWS" while firing the command "kops create cluster --name legend.k8s.local --zones ap-south-1a --master-size t2.micro --node-size t2.micro" can you please help here I m stuck
Hi i wanna ask something, did you create hosted zone before running kops ? is it required to create hosten zone in route53 before start creating cluster with kops ? Thanks for the video btw, i learn so much from this channel
Hi, thanks for watching. Hosted zone and route53 entry is required. But if you chose the cluster name same as I shown in this video, then you don't need to do the hosted zone/route 53 bit.
@@justmeandopensource thanks for replying, i just search for that and i found this stackoverflow.com/questions/55049641/kubernetes-kops-without-dns thankyou for your awesome works, currently finishing your kubernetes video series, keep it up!
Hi Venkat, Have a question about deploying Kubernetes. we are Ansible shop (our datastack is deployed using Ansible on-Prem. we want to deploy Kubernetes on-prem (on VM's) my question is whats the best way to deploy K8s -- I see options like Kubeadmin, Kops, Kubespray etc
HI Manoj, thanks for watching. I have all provisioning related videos grouped in a separated playlist. th-cam.com/play/PL34sAs7_26wODP4j6owN-36Vg-KbACgkT.html Kubeadm is entirely manual. KOPS is designed for Cloud (AWS mainly but beta support for GCP and Azure). You mentioned you want to deploy on-prem. I would choose Kubespray as its based on Ansible which you are already using. Using Kubespray you can provision production ready cluster with High availability. Managing the cluster is also simple with Kubespray. th-cam.com/play/PL34sAs7_26wOAqYsrIhtDaIviGlSkmfv9.html Cheers.
hello i have a small query here , does Aws use bastion/jumpbox in terms of giving us control to our cluster resources created in Aws. when we use our kops cluster by running kops or by our local system by accessing the .kube/config in your local system it give access to cluster coz there are private/public IP of aws instance when we check the IP address manually we get private but when we check on Aws console its public. Just a small doubt, suggestions please !
Hi, thanks for watching. There is no Bastion or Jump host involved here. You created all the instances yourself where you wanted to deploy Kubernetes through KOPS automation. So you have full control to who can access your instances and so on. You must have configured the security groups. KOPS does provision the cluster but you will have to make sure its exposed to the right network for access. Cheers.
@@justmeandopensource oh you mean to say we just have to configure the security group as per our need in terms to access the cluster, am confused coz when we create instance on aws and install k8s on it and after copying the .kube/config to our local system we cant make connection to the cluster coz their it reaches for the private ip not the public. This is my confusion.
@@justmeandopensource oh I see.... Thankyou so very much will try this now.... Well I been to many videos and took many courses but believe me you are really amazing. Soon going to give interviews and believe me all credit will go to you for sharing and making things understanding easily. Thanks !! Cheers.
Sorry, what is the default user for debian machine when it comes to windows machine while the master and worker machines were created by using kubernetes kops server? Hope i will get reply soon
Hi Adrian, thanks for watching. KOPS creates load balancer to load balance between multiple master/controller nodes. I haven't paid much attention to the type of load balancer it deploys. But I would like to understand what problem do you have with classic load balancer that you want to switch to NLB or ALB. Classic load balancer is more than sufficient in this case. If you are trying to use the same load balancer for your applications deployed within the cluster, then that might not be the right way.
@@justmeandopensource Hi thank you for replying. To be clear, I am planning on attaching an existing ALB or NLB to my worker nodes where my applications are running.
Hello Venkat! Thanks for the amazing video. I have a question, if we have to export the Kube config on worker node , so that we can run kubectl commands there, how do we do that on Kops?
Hi. First of all, I'd like to thank you for all your amazing videos! Second, is there a way to create the root volumes of master and worker nodes as encrypted volumes via kops? I can only find encryption for etcd volumes for master, and additional volumes for worker nodes. Thanks!
Hi Venkat.. thanks for the videos ..your videos are a great source for learning k8s.. i am trying to create a cluster with same .. but load banner instance are out of service always in hc .. not able to validate cluster
Hi thanks for the video. Instead of classic load balancer how can I create Application loadbalancer through kops. For security reasons classic loadbalancer is not allowed. Can you help with this creating Application loadbalancer through kops.
Hi, thanks for watching. I use Arch Linux as the base OS and Termite as the terminal emulator with other goodies. I have done a video on my terminal setup recently if you would like to watch it th-cam.com/video/PUWnCbr9cN8/w-d-xo.html
Hi VenKat, Thank you showing us the value knowledge via this channel. I have an error showing below. Do I need to use "sshkeygen..." to generate the key first then try again? If yes, if you have any exactly command for that , please provide here. Appreciation for your help: " SSH public key must be specified when running with AWS (create with `kops create secret --name justme.k8s.local sshpublickey admin -i ~/.ssh/id_rsa.pub`)"
got another issue (cannot create pod nginx) after temporary pass the sshpublic key like below: "Plugin kubenet failed to set up pod "nginx-6db489d4b7-mwzpb_default" network: error adding container to network: invalid range set 0: Network has host bits set. For a subnet mask of length 24 the network address is 100.96.1.0 " the pod container related to kube-dns and dns-autoscaler not working "deployment.apps/kube-dns 0/1 1 0 31m
Hi Venkat, this is an awesome course. I was just curious if 1) i want to create the cluster in private subnet and create a tunnel to it through a public subnet then how would you recommend we should do. 2) how would i automate this infrastructure creation/update. Hope you can do a course on these.
Thank you so much for the amazing videos. I am trying to learn the difference between kops and kubespray. Currently I am using Kubespray for microservice deployment however is kops better than kubespray ? Is Kops mostly used on public cloud ?
Hi Sherin, thanks for watching. Both Kops and Kubespray are mature and production ready. While Kubespray is designed to be used on any insfratructure, KOPS is built primarily for AWS but it also support other cloud providers.
@@justmeandopensource Thank you for clarifying. So asking further on this: Is upgrading in kubespray cumbersome than kops ?When it comes to managing clusters which is more preferred ?
@@sherinrajan8220 I have tried both but haven't used enough to be able to comment on which one is better. It all depends on your use case. The below medium blog might help you. medium.com/@m.k.joerg/overview-of-kubernetes-installers-8f06437d215a
Hey, you have made great video again. I just wanted to ask - is there any way to make cluster where I deploy stateful deployment on spot instances, and if spot instance not available then automatically it acquires on-demand?? Thanks in advance. 😀
Hi Anushman, thanks for watching. you will have to have your worker nodes provisioned and deployed before you can deploy any resources like statefulsets. Its not like when you deploy a statefulset, it will launch a spot instance and then get deployed on it.
@@anshuman2121 The worker nodes are provisioned in node groups. In node groups you specify what instance type, minimum count and maximum count. If a node gets deleted, the node group will make sure to have the minimum count of worker nodes. I haven't explored spot instances yet in node groups.
Hi venkat. After installing the kops server, we have created master and worker machines in the kops server. Already we have created kubectl in kops server. My doubt is can i again install any softwares in master and worker machines for to integrate jenkins server with kubernetes? Please help me.
Hi Kotapati, thanks for watching. You mentioned "After installing kops server". Kops is not a server. Its a tool/binary. I guess thats what you meant. Again you mentioned "we have created master and worker nodes in the kops server". There is no kops server. I assume you provisioned the cluster using kops command in AWS. Kops is just for provisioning the cluster. You then have ssh access to all the nodes including the master nodes. You can install any software on them as you like. But its not advisable to alter any of the kubernetes components as they are effectively and well managed by the kops tool.
Hi, I replied to your other comment about this. Anyways here is a video of mine that might answer most of your questions. th-cam.com/video/PUWnCbr9cN8/w-d-xo.html&pp=ygUQanVzdCBtZSB0ZXJtaW5hbA%3D%3D
Hi Praveen, thanks for watching. Starting at 2:15 in this video, I have shown how to download the binary from github release page, then set the executable permission and move it to /usr/local/bin. Did you miss any of the above steps? Did you actually download the kops binary?
Hi, thanks for watching. I use ZSH shell on any terminal emulator with some plugins like zsh-autosuggestions and zsh-syntax-highlighting. I have done a video on my terminal setup if you would like to watch. Here it is th-cam.com/video/PUWnCbr9cN8/w-d-xo.html&pp=ygUQanVzdCBtZSB0ZXJtaW5hbA%3D%3D
Hi Venkat, this is a very informative video on KOPS Cluster. I am trying to create KOPS K8s cluster using Terraform (generated by KOPS) and getting following error. Can you please provide some pointers so as to get unblocked from this error? aws_route53_zone_association.diy-kops-k8s-cluster: Still creating... [1m0s elapsed] aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Still creating... [50s elapsed] aws_route53_zone_association.diy-kops-k8s-cluster: Still creating... [1m10s elapsed] aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Still creating... [1m0s elapsed] aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Still creating... [1m10s elapsed] aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Creation complete after 1m14s [id=nodes.diy.kops.k8s.cluster] Error: error getting Route 53 Zone Association (/hostedzone/Z0123456Z3G6T410IE:vpc-0f1aa123456:us-east-1): missing after creation I checked with my R53 Hosted Zone and i can see that its associated with the above listed VPC.
Hi Venkat, thanks for watching. I will explore that. When creating a cluster, you can specify the type of overlay network instead of going for the default. But haven't tested changing the network type once deployed. Will have to do some testing. Cheers.
Hi, thanks for watching. There are options to choose an existing subnet when deploying cluster through KOPS. I have various topics that I need to cover. I will try to come back to this at some point. Cheers.
Great channel but I was wondering how I could integrate Terraform with kops . Documentation states to add --out=. \ --target=terraform \ flags to kops create command -- but I keep getting a read-only file system error. Error: error closing target: error creating output directory "data": mkdir data: read-only file system. Has anyone encountered and resolved this issue?
Hi Mayur, thanks for watching and suggesting Patreon. I didn't create a Patreon account as I can't commit more time than I have now. Also not wanted to monetize my knowledge. Just sharing out of passion. Cheers.
What kind of load balancer you have deployed ? Is it nginx ingress controller ? When i create cluster it does not create load-balancer with it. Let me show you the command. kops create cluster \ --state=${KOPS_STATE_STORE} \ --node-count=2 \ --master-size=t2.micro \ --node-size=t2.micro \ --zones=ap-south-1a,ap-south-1b \ --name=${KOPS_CLUSTER_NAME} \ --dns public \ --master-count 1
HI Sarfaraz, thanks for watching. Nginx ingress controller is not used as a load balancer but as a proxy server. In this video, kops create an AWS load balancer for the master nodes. You can see that in Load Balancer section in EC2 service in your AWS management console. Thanks
Also the DNS I used specifically is the one recommended for a quick start testing. If you use anyother DNS it will fail unless the DNS is a valid public one.
Great Intro, thank you. We are using kops toolbox to create clusters, would be good if you can create a small video on using the same and highlight on components like etcd members, clusters and other resources. Some prelim info is here github.com/kubernetes/kops/blob/master/docs/cluster_spec.md but not a very detailed one.
Underrated channel, a good practice from Just me and Opensource
Such a great explanation & fluently understandable.
Kudos to you Mr Venkat Nagappan
Hi Muzaffar, many thanks.
Dude. You're such a genius. Don't know how I spent the last 33 mins watching this video. I love the way you explain the concepts with such great detail. This is my first video of your channel. I am already a great fan of yours. Keep up the spirit and spread the knowledge for free.
Hi Samarth, many thanks for your interest in my content. Cheers.
@@justmeandopensource First time subbie here. Just discovered your content tonight and I'm wondering why TH-cam Algo hasn't done the needful since. You're so talented, Bro. Will be recommending and following up all your stuff.
Man you are an amazing human being. You made the KOPS methods as easy as ABCD. WOW!!!
Hi Nana, many thanks for watching.
Nice Explanation which is related to KOPS; Thanks bro
😀
Glad you liked it
it is great tutorial. it is very useful for me, thanks
I was looking for kops templating to create cluster, landed on this video.
Thanks for watching.
Thank you very much for your whole video series, very concise and informative. I learnt so much in less than 1hr.
Thanks for watching.
Already waiting for the next installment in this series ! thanks a lot ...
Hi Cokegen, thanks for watching. Cheers.
@@justmeandopensource I was thinking on the upgrade process and it just so happens that I have to upgrade a cluster from 1.15.x to 1.16.x, and an important piece of that upgrade is the CNI part, in my case, maybe going from a somewhat ancient version of Calico to a newer one. Not saying that you cover exactly my use case, but if you can cover how to maintain things working WHILE also upgrading the CNI portion of the cluster, I guess that would be really really useful and it's a real situation on lots of production systems. Thanks, loving the content on your channel ...
Let me explore that scenario.
I am really thankful that you made this video. God bless you!!!
Hi Manmohan, thanks for watching.
You are genius, this is really helpful.
Thanks for watching.
Great work, keep the videos coming!!!
Hi, thanks for watching. Cheers.
This is what i have been looking from your channel, please cover all the k8s topics with aws
Will do. Thanks for watching. Cheers.
Appreciate your time and knowledge sharing
Hi, Thanks for watching.
Very well said thank you
Thanks for watching.
Just started watching your Kubernetes #limitless time
Thank you. It very useful video.
Hi, thanks for watching.
Please make a video on how to use the eks distro with kOps on AWS
Nice one :)
b/w which OS distro are you using?.
Arch Linux with I3 tiling window manager. Thanks for watching.
Can you please let us know which terminal is using?
super👌👌🔥🔥plz post ci/CD pipelines..
Hi Jai, thanks for watching.
Hi!
FoA, thanks for a video! Can u tell me: how a u did autofill cli? (show tips when you write command)
Hi, thats by the use of Zsh with some cool plugins. I have explained my setup in this video.
th-cam.com/video/PUWnCbr9cN8/w-d-xo.html
I have given image id, then it start working. Very good video. Thank you. command given : kops create cluster --name justme.k8s.local --zones eu-west-1a --master-size t2.micro --node-size t2.micro --image "ami-0eb11ab33f229b26c" --kubernetes-version 1.15.0
do you have to register a domain name for the sub domains in your cluster ? I have not see in your tutorial .. thanks --James
Hi James, thanks for watching. Apologies for not clarifying about DNS in this video.
I didn't register any domain instead used gossip-based domain.
Basically anything ending with k8s.local. In this video I used justme.k8s.local
kops.sigs.k8s.io/getting_started/aws/#prepare-local-environment
@@justmeandopensource - Thank you for your quick reply .. more power to you
@@jamesbonding3988 No worries. You are welcome.
which editor you are using for writing these commands
Hi Ramen, thanks for watching. Do you mean the terminal editor? Its just gnome-terminal
Quick question:
so after creating cluster using kops tool why can't i find the cluster under EKS service in aws account.
Kops doesn't create AWS EKS managed kubernetes cluster but just set of ec2 instances and kubernetes on top of it. So you won't see the cluster in EKS console. Just check the ec2 instances.
Hi Venkat, Great channel! Forgive my ignorance, but for the high availability (HA) masters set-up, how does KOPS make it possible to use a single ELB Classic load balancer to serve traffic to two different Auto Scaling Groups each with their own master instances? I didn't think this was possible unless an ALB was used. What am I missing in my understanding? Thanks!
Thanks venkat for valuable session.
What is the default user for ubuntu machine when it comes to windows machine?
Hi, Thanks for watching. If you used the default ami, then the user is admin. You can see that I am logging in at 29:20 in this video.
Thanks for this video..I'am new to devops and want to switch job in devops position, but its getting very difficult for me to prepare for technical interviews since I dont have real time experience of devops culture and work. I would be very grateful to you if you would provide some overview of how to prepare for devops technical interview.What should be the preparation areas..topics to cover..not the technicalities or technical details but the overall idea of how should be the preparation points and topics. It would be a much help.
Many many Thankyou in advance.
Hi venkat
Thanks for this video....
Can you please post a video for kubernetes networking (cni weave)or anything
Hi Surendar, thanks for watching. In all my Kubernetes provisioning videos, I used either flannel or Calico as the overlay network. Installing Weave as CNI is very similar. Apart from installing the overlay network, I haven't looked much into the features of inidividual providers. Soon I will be exploring pod network policies which requires Calico or Weave as the overlay network. Cheers.
Hello Venkat,
Thank you for posting this video.
I was trying to implement kops following your step-by-step video but couldn't connect to the master and nodes via ssh. It throws a "permission denied(publickey) error.
How do I fix this as I didn't specify the key while creating the cluster? is it possible to generate a key and apply it after the cluster has been already created?
Hi, could you please tell me when we create a cluster (master and nodes) using kops it's default Os is ubuntu or something else. so my question is how can we give a specific Linux Os for both master and nods excepts default os. I mean can we change the default os as a cento or Red Hat Linux ? .thanx
Hi KLP, thank for watching. By default, if you set up a cluster using KOPS as of today, the default OS will be Ubuntu 20.04. KOPS supports a lot of Linux distribution.
Take a look at the below documentation on how to change the OS. You basically have to create your own AMI with one of the supported OS and the edit kops instance group to use a particular image.
github.com/kubernetes/kops/blob/ed32f19591769a712abfb2fb52f37adc5f3aba44/docs/operations/images.md
Cheers.
@@justmeandopensource Thank you.
@@lakshithapromodaya You are welcome. Cheers.
Hi, could you please explain what will happen if the master node got terminated? I had an experienced, my master node terminated and
recreated it self help of an autoscaling group.but after recreated a new master node, my cluster is not worked
perfectly like previous(FYI: I used AWS dynamic Ip address. after the recreated master node IP address also changed)
Hi KLP, thanks for watching. I haven't tested that scenario. The node instance group is responsible for maintaining the specified number of nodes in that group. So if one of your master died/crashed, the instance group will bring up another one but I believe you will have to run KOPS again to set up that new master node. I am not entirely sure about it.
@@justmeandopensource Thanks..
@@lakshithapromodaya You are welcome. Cheers.
Hi Venkat, Is it possible for u to explain dual IP stack support in kubernetes. Maybe some demo about how to enable the same and also validate if a pod has both IP addresses.
Hi Amartya, thanks for watching. I have read about the dual stack but have never explored it. I have videos for the next two months scheduled. Meanwhile when I get a chance I will explore this. Cheers.
Hi Venkat. thanks much . It’s magic for creating k8s cluster now using kops.
I tried creating cluster with private topology and bastion. And the cluster is ready. However I am not sure how to do ssh to the master and worker nodes as they are not having public up. Pls assist. Thanks.
Hi, thanks for watching. I haven't tried that but did you use the option --ssh-public-key when creating the cluster using kops create cluster command? You may have already seen this article but if you haven't,
medium.com/andcloudio/kubernetes-kops-cluster-on-aws-f55d197d8304
Hello Venkat, I followed the above steps on my newly created AWS account , I am getting the error "SSH public key must be specified when running with AWS" while firing the command "kops create cluster --name legend.k8s.local --zones ap-south-1a --master-size t2.micro --node-size t2.micro" can you please help here I m stuck
Hi i wanna ask something, did you create hosted zone before running kops ? is it required to create hosten zone in route53 before start creating cluster with kops ?
Thanks for the video btw, i learn so much from this channel
Hi, thanks for watching. Hosted zone and route53 entry is required. But if you chose the cluster name same as I shown in this video, then you don't need to do the hosted zone/route 53 bit.
@@justmeandopensource thanks for replying, i just search for that and i found this stackoverflow.com/questions/55049641/kubernetes-kops-without-dns
thankyou for your awesome works, currently finishing your kubernetes video series, keep it up!
@@fhmisml Cool.
Hi Venkat, Have a question about deploying Kubernetes. we are Ansible shop (our datastack is deployed using Ansible on-Prem. we want to deploy Kubernetes on-prem (on VM's) my question is whats the best way to deploy K8s --
I see options like Kubeadmin, Kops, Kubespray etc
HI Manoj, thanks for watching. I have all provisioning related videos grouped in a separated playlist.
th-cam.com/play/PL34sAs7_26wODP4j6owN-36Vg-KbACgkT.html
Kubeadm is entirely manual. KOPS is designed for Cloud (AWS mainly but beta support for GCP and Azure). You mentioned you want to deploy on-prem. I would choose Kubespray as its based on Ansible which you are already using. Using Kubespray you can provision production ready cluster with High availability. Managing the cluster is also simple with Kubespray.
th-cam.com/play/PL34sAs7_26wOAqYsrIhtDaIviGlSkmfv9.html
Cheers.
hello i have a small query here , does Aws use bastion/jumpbox in terms of giving us control to our cluster resources created in Aws. when we use our kops cluster by running kops or by our local system by accessing the .kube/config in your local system it give access to cluster coz there are private/public IP of aws instance when we check the IP address manually we get private but when we check on Aws console its public. Just a small doubt, suggestions please !
Hi, thanks for watching. There is no Bastion or Jump host involved here. You created all the instances yourself where you wanted to deploy Kubernetes through KOPS automation. So you have full control to who can access your instances and so on. You must have configured the security groups. KOPS does provision the cluster but you will have to make sure its exposed to the right network for access. Cheers.
@@justmeandopensource oh you mean to say we just have to configure the security group as per our need in terms to access the cluster, am confused coz when we create instance on aws and install k8s on it and after copying the .kube/config to our local system we cant make connection to the cluster coz their it reaches for the private ip not the public. This is my confusion.
@@realthought2262 In the downloaded kube config file you can change the ip address to that of your public ip address.
@@justmeandopensource oh I see.... Thankyou so very much will try this now.... Well I been to many videos and took many courses but believe me you are really amazing. Soon going to give interviews and believe me all credit will go to you for sharing and making things understanding easily. Thanks !! Cheers.
Thanks for your kind words. People like you keep me motivated to do more of what I love doing. Cheers.
Thanks for the content Venkat. Can you please help in providing some steps to create cluster in eks using kubeconfig
Sorry, what is the default user for debian machine when it comes to windows machine while the master and worker machines were created by using kubernetes kops server?
Hope i will get reply soon
its admin
Hello, I have a question, kops creates the load balancer but by default it's a classic load balancer. Is it possible to change this to an ALB or NLB?
Hi Adrian, thanks for watching. KOPS creates load balancer to load balance between multiple master/controller nodes. I haven't paid much attention to the type of load balancer it deploys. But I would like to understand what problem do you have with classic load balancer that you want to switch to NLB or ALB. Classic load balancer is more than sufficient in this case. If you are trying to use the same load balancer for your applications deployed within the cluster, then that might not be the right way.
@@justmeandopensource Hi thank you for replying. To be clear, I am planning on attaching an existing ALB or NLB to my worker nodes where my applications are running.
@@MrAjsanpedro The load balancer KOPS created is for the master nodes. You can attach your existing ALB to worker nodes manually as you like.
Hello Venkat!
Thanks for the amazing video. I have a question, if we have to export the Kube config on worker node , so that we can run kubectl commands there, how do we do that on Kops?
Hi. First of all, I'd like to thank you for all your amazing videos!
Second, is there a way to create the root volumes of master and worker nodes as encrypted volumes via kops? I can only find encryption for etcd volumes for master, and additional volumes for worker nodes. Thanks!
How do you get that auto-complete within the terminal?
Hi Jason, I did a video on my terminal setup.
th-cam.com/video/PUWnCbr9cN8/w-d-xo.html
Hi Venkat.. thanks for the videos ..your videos are a great source for learning k8s.. i am trying to create a cluster with same .. but load banner instance are out of service always in hc .. not able to validate cluster
Hi thanks for the video. Instead of classic load balancer how can I create Application loadbalancer through kops. For security reasons classic loadbalancer is not allowed. Can you help with this creating Application loadbalancer through kops.
Hi Ashwath, thanks for watching. I will look into this. I haven't really paid much attention to the type of load balancer.
Bro could you please do video which is the best method to take backups and how do we do restore if any deletion happened. I am looking for it.
what terminal are you using?
Hi, thanks for watching.
I use Arch Linux as the base OS and Termite as the terminal emulator with other goodies.
I have done a video on my terminal setup recently if you would like to watch it
th-cam.com/video/PUWnCbr9cN8/w-d-xo.html
Hi VenKat,
Thank you showing us the value knowledge via this channel. I have an error showing below. Do I need to use "sshkeygen..." to generate the key first then try again? If yes, if you have any exactly command for that , please provide here. Appreciation for your help:
" SSH public key must be specified when running with AWS (create with `kops create secret --name justme.k8s.local sshpublickey admin -i ~/.ssh/id_rsa.pub`)"
got another issue (cannot create pod nginx) after temporary pass the sshpublic key like below:
"Plugin kubenet failed to set up pod "nginx-6db489d4b7-mwzpb_default" network: error adding container to network: invalid range set 0: Network has host bits set. For a subnet mask of length 24 the network address is 100.96.1.0
"
the pod container related to kube-dns and dns-autoscaler not working
"deployment.apps/kube-dns 0/1 1 0 31m
Does kubectl and kops must be in same version?
No. But keep the kubectl version +/- one version around the kubernetes version that kops deploys. Hope it makes sense. Thanks for watching.
Hi Venkat, this is an awesome course. I was just curious if
1) i want to create the cluster in private subnet and create a tunnel to it through a public subnet then how would you recommend we should do.
2) how would i automate this infrastructure creation/update.
Hope you can do a course on these.
Thank you so much for the amazing videos. I am trying to learn the difference between kops and kubespray. Currently I am using Kubespray for microservice deployment however is kops better than kubespray ? Is Kops mostly used on public cloud ?
Hi Sherin, thanks for watching. Both Kops and Kubespray are mature and production ready. While Kubespray is designed to be used on any insfratructure, KOPS is built primarily for AWS but it also support other cloud providers.
@@justmeandopensource Thank you for clarifying. So asking further on this: Is upgrading in kubespray cumbersome than kops ?When it comes to managing clusters which is more preferred ?
@@sherinrajan8220 I have tried both but haven't used enough to be able to comment on which one is better. It all depends on your use case. The below medium blog might help you.
medium.com/@m.k.joerg/overview-of-kubernetes-installers-8f06437d215a
@@justmeandopensource Thank you.
is it possible to use kops without aws directly on bare metal.
Hi, thanks for the video, can you explain about gossip based cluster as there is not much info in the documentation.
Thanks mate!!
Hey, you have made great video again.
I just wanted to ask - is there any way to make cluster where I deploy stateful deployment on spot instances, and if spot instance not available then automatically it acquires on-demand?? Thanks in advance. 😀
Hi Anushman, thanks for watching. you will have to have your worker nodes provisioned and deployed before you can deploy any resources like statefulsets. Its not like when you deploy a statefulset, it will launch a spot instance and then get deployed on it.
@@justmeandopensource Can we automate the things like if any spot instance goes away then it on-demand instance automatic provision?
And automatically reclaim the spot instance
@@anshuman2121 The worker nodes are provisioned in node groups. In node groups you specify what instance type, minimum count and maximum count. If a node gets deleted, the node group will make sure to have the minimum count of worker nodes. I haven't explored spot instances yet in node groups.
I love unicorns :)
Hi Jason, thanks for noticing that.
Hi venkat. After installing the kops server, we have created master and worker machines in the kops server. Already we have created kubectl in kops server.
My doubt is can i again install any softwares in master and worker machines for to integrate jenkins server with kubernetes?
Please help me.
Hi Kotapati, thanks for watching.
You mentioned "After installing kops server". Kops is not a server. Its a tool/binary. I guess thats what you meant. Again you mentioned "we have created master and worker nodes in the kops server". There is no kops server. I assume you provisioned the cluster using kops command in AWS.
Kops is just for provisioning the cluster. You then have ssh access to all the nodes including the master nodes. You can install any software on them as you like. But its not advisable to alter any of the kubernetes components as they are effectively and well managed by the kops tool.
Hi, Is it possible to share the terminal that you used and how this auto completion works. it would be helpful to know
Hi, I replied to your other comment about this. Anyways here is a video of mine that might answer most of your questions. th-cam.com/video/PUWnCbr9cN8/w-d-xo.html&pp=ygUQanVzdCBtZSB0ZXJtaW5hbA%3D%3D
What would you say are the disadvantages of using kops? or is it only just advantages?
Hi ..Venkat,
I tried for the same accordingly you but its trowing error when i type which /usr/local/bin/kops no user and path
Hi Praveen, thanks for watching. Starting at 2:15 in this video, I have shown how to download the binary from github release page, then set the executable permission and move it to /usr/local/bin.
Did you miss any of the above steps? Did you actually download the kops binary?
yes resolve it thanks.
after running kubectl get nodes --asking for id pass? .what is the id pass ? .please tell me
Hi Venkat, thanks for the video.
I am getting one error inside my kops cluster which is, I am unable to access to internet from the pods.
can i know the terminal used and how auto complete works
Hi, thanks for watching. I use ZSH shell on any terminal emulator with some plugins like zsh-autosuggestions and zsh-syntax-highlighting. I have done a video on my terminal setup if you would like to watch. Here it is th-cam.com/video/PUWnCbr9cN8/w-d-xo.html&pp=ygUQanVzdCBtZSB0ZXJtaW5hbA%3D%3D
Thankyou it works@@justmeandopensource
@@UndecillionTechbyte no worries
Hi Venkat, this is a very informative video on KOPS Cluster. I am trying to create KOPS K8s cluster using Terraform (generated by KOPS) and getting following error. Can you please provide some pointers so as to get unblocked from this error?
aws_route53_zone_association.diy-kops-k8s-cluster: Still creating... [1m0s elapsed]
aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Still creating... [50s elapsed]
aws_route53_zone_association.diy-kops-k8s-cluster: Still creating... [1m10s elapsed]
aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Still creating... [1m0s elapsed]
aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Still creating... [1m10s elapsed]
aws_autoscaling_group.nodes-diy-kops-k8s-cluster: Creation complete after 1m14s [id=nodes.diy.kops.k8s.cluster]
Error: error getting Route 53 Zone Association (/hostedzone/Z0123456Z3G6T410IE:vpc-0f1aa123456:us-east-1): missing after creation
I checked with my R53 Hosted Zone and i can see that its associated with the above listed VPC.
can you please make a video on upgrade of calico using kops edit cluster if possible
Hi Venkat, thanks for watching. I will explore that. When creating a cluster, you can specify the type of overlay network instead of going for the default. But haven't tested changing the network type once deployed. Will have to do some testing. Cheers.
Hi bro, can you create a video on How to install kops cluster inside private subnet
Hi, thanks for watching. There are options to choose an existing subnet when deploying cluster through KOPS. I have various topics that I need to cover. I will try to come back to this at some point. Cheers.
Great channel but I was wondering how I could integrate Terraform with kops . Documentation states to add --out=. \
--target=terraform \ flags to kops create command -- but I keep getting a read-only file system error. Error: error closing target: error creating output directory "data": mkdir data: read-only file system. Has anyone encountered and resolved this issue?
Always loves your Videos and Thank you for sharing with us. Do you have your Patrons account? so we can contribute you and you can help us even more.
Hi Mayur, thanks for watching and suggesting Patreon. I didn't create a Patreon account as I can't commit more time than I have now. Also not wanted to monetize my knowledge. Just sharing out of passion. Cheers.
@@justmeandopensource Respect +1
@@mayurchavhan8590 Cheers.
What kind of load balancer you have deployed ? Is it nginx ingress controller ? When i create cluster it does not create load-balancer with it. Let me show you the command. kops create cluster \
--state=${KOPS_STATE_STORE} \
--node-count=2 \
--master-size=t2.micro \
--node-size=t2.micro \
--zones=ap-south-1a,ap-south-1b \
--name=${KOPS_CLUSTER_NAME} \
--dns public \
--master-count 1
HI Sarfaraz, thanks for watching. Nginx ingress controller is not used as a load balancer but as a proxy server. In this video, kops create an AWS load balancer for the master nodes. You can see that in Load Balancer section in EC2 service in your AWS management console. Thanks
I followed the same steps but my load balancer was not created , what did i miss?
Hi, thanks for watching. Do you consistently get this problem every time?
can you share with me the name of the terminal
Its gnome Terminal on Manjaro Linux with I3 tiling window manager.
Tnq sir..subscribed
How to contact.. please tell
I don't know how his cluster get created without ssh key without dns. or he forgot to tell about that.
You don't need an ssh key unless you want the ability to log in the cluster nodes later.
Also the DNS I used specifically is the one recommended for a quick start testing. If you use anyother DNS it will fail unless the DNS is a valid public one.
Great Intro, thank you. We are using kops toolbox to create clusters, would be good if you can create a small video on using the same and highlight on components like etcd members, clusters and other resources. Some prelim info is here github.com/kubernetes/kops/blob/master/docs/cluster_spec.md but not a very detailed one.
Hi Arpan, thanks for your comment. I will look into it. Cheers.