When trying to run aws eks update-config I got the error "'NoneType' object is not iterable". I think it's a bug from AWS CLI when you already have a config file on your machine. But you can get a workaround by renaming the config file or deleting it. (rm ~/.kube/config). Hope it helps.
I would love to see a tutorial how to create a Kubernetes deploy job so that I could manually deploy a specific build to a specific service and specific environment. For instance, after I've pushed my new branch to GitHub and the build is finished I would like to be able to deploy that build to a specific service in a specific environment (staging, dev, pre-flight, production, etc.) in order to show my customer what I've done.
Hi Marcel..good day..I am a new subscriber and going to watch all the videos and learn Devops technology. I learnt Devops from class room training close to one year..but I didnt get any job.I am from BPO background. Working hard to achieve in Devops.. thank you that videos. Please prepare videos for any scripting language for automation. 😍👍
request... please cover how to install node_exporter and/or *Beats modules onto managed node_groups. - aka expose the metrics and logs to Prometheus and EK stack.
check out these two videos. Both cover node exporter and are cloud agnostic czcams.com/video/1-tRiThpFrY/video.html czcams.com/video/abq-6FBn0fo/video.html
@@MarcelDempers hi hi, looked, it covers nicely how to deploy prometheus and node-exporter, but not how to pin the deployment onto a selected node_group. What I got is a single control plane, and then 3 node groups, 1. app, 2. db, 3.management. I want to create the monitoring namespace and pin it onto the management node_group. and as this is EKS, PodNodeSelector is not a option, due to size of the entire stack it's not practical to add a nodeSelector to each pod deployment yaml file... to many files to edit, to much chance for mistakes...
@@MarcelDempers ... thinking... I can do a complete directory search for : nodeSelector and then add the relevant selector to all the code, might a option... a nicer option would of course be if I could create a namespace and pit it to a node_group
@@georgelza It's often that folks want to patch a deployment without having to edit all the files. Take a look at `kubectl patch -n namespace deployment mydeployment --patch $(cat patch.yaml) This allows you to bundle a set of patches along with all your templates In patch.yaml, you add only the node selector spec: template: spec: nodeSelector: foo: bar
@@MarcelDempers my next ask is to get persistent volume added to the script. wanted to use EFS, but been informed Prometheus is not approved for EFS, has to be EBS.
Hi Mercel, I'm following your step, but when I'm trying to download kubectl from google I'm getting *The specified key does not exist*, do you know if the url changed? Thanks. Awesome tutorial.
Just checking quickly, when you say use ekscli... and create the control plane... my VPC already have 2 subnets in 2 AZ's... we know for the control plane, you actually really want it deployed across 3 AZ's. so question, as the control plane is a managed service, does it get create inside my VPC, (other words I need to have 3 subnets) or is it done outside my VPC, and all that goes into my VPC is the EC2 instance (which implies I can deploy them across my 2 configured AZ's)
I can't recall off the top of my head but as far as I'm aware, a managed cluster does not rely on your subnets as the control plane will run on Amazon's network and not your VPC.
@@MarcelDempers My thinking also, but seem to remember you providing subnet names and security group names when you created the management plane, let me look at your video again.
Thanks for the video. I created role and attached policies for the nodes but, I get this error while creating the node group. "An error occurred (InvalidRequestException) when calling the CreateNodegroup operation: Failed to create service linked role: AWSServiceRoleForAmazonEKSNodegroup due to missing permissions for 'iam:CreateServiceLinkedRole' "
Hi Marcel, thanks for the video. I have a question, when we create our EKS cluster (through CLI or web console), does AWS create a separate VPC to host the control plane while attaching resources and ENI to our specified VPC and subnets? Thank you!
I would not really consider either as best practise. They are both tools that solve different things. aws-cli helps you talk to AWS to provision resources. It's a key to understand aws-cli when working with AWS. Terraform helps provision infrastructure as code and can help manage infra at bigger scale, so it's also important to learn. It's always best to understand aws-cli as an AWS user first, then take a look at Terraform to see what it solves for you.
i keep getting this error when defining worker node policy aws iam create-role --role-name getting-started-eks-role-nodes --assume-role-policy-document .assume-node-policy.json An error occurred (MalformedPolicyDocument) when calling the CreateRole operation: This policy contains invalid Json
To follow the series, feel free to bookmark my "Kubernetes in the Cloud" playlist --> czcams.com/play/PLHq1uqvAteVsUhzNBkn-rPzXtPNpJu1-k.html
your tutorial cures my depression, thanks
This is amazing, you're so clear compared to every other tutorial on kuber I've seen... bravo
When I feel like learning something new, I search for that devops guy!
I just subscribed today and my first thought was: "Why haven't I subscribed earlier?"
Wow!! You explain it so well in such a simple manner
This is fantastic stuff. I will be a lifetime follower it seems. Things are explained so well and in modular fashion. Thanks a ton. Keep them coming!
I blame myself for not finding you earlier to this! Glad I am following you at least now! Great job!!
When trying to run aws eks update-config I got the error "'NoneType' object is not iterable". I think it's a bug from AWS CLI when you already have a config file on your machine. But you can get a workaround by renaming the config file or deleting it. (rm ~/.kube/config). Hope it helps.
Thank you soooooo much, everything we need to know in 30 minutes, that is a very good job !
Amazing stuff, very clear and to the point.
@That DevOps Guy ; Great content. Any chance for GKE videos ? especially utilising cloud sql sidecart pattern, it's giving me a headache.
Great, efficient, to-the-point intro to AWS EKS! Thanks!
great video! the source code are missing in git. You got a new fan
Thanks so much Marcel, really appreciate the content
I would love to see a tutorial how to create a Kubernetes deploy job so that I could manually deploy a specific build to a specific service and specific environment. For instance, after I've pushed my new branch to GitHub and the build is finished I would like to be able to deploy that build to a specific service in a specific environment (staging, dev, pre-flight, production, etc.) in order to show my customer what I've done.
Thanks for sharing. Save my time. ^^
Hi, I have a question, if I'm using EKS CLI, should I create policies, roles, and everything you show before to AWS CLI? Thanks for your videos.
With the eksctl, the inputs are simple, you dont need to create all the policies
oles etc upfront. Those are only required with awscli 💪🏽
Hi Marcel..good day..I am a new subscriber and going to watch all the videos and learn Devops technology. I learnt Devops from class room training close to one year..but I didnt get any job.I am from BPO background. Working hard to achieve in Devops.. thank you that videos. Please prepare videos for any scripting language for automation. 😍👍
What if it is not on free tier could loft be a solution to prevent him cost?
Marcel, Thumbs up. All your videos are well explained. Thank you. Keep up the good work.
Great tutorial. Wish I'd have found it before I spent hours trying to sort out deploying kubernetes. Azure was way easier.
awesome!!! Thank you Marcel. Great content.
Nice one, please make a video about using terraform, kops to create AWS EKS cluster.
request... please cover how to install node_exporter and/or *Beats modules onto managed node_groups. - aka expose the metrics and logs to Prometheus and EK stack.
check out these two videos. Both cover node exporter and are cloud agnostic
czcams.com/video/1-tRiThpFrY/video.html
czcams.com/video/abq-6FBn0fo/video.html
@@MarcelDempers hi hi, looked, it covers nicely how to deploy prometheus and node-exporter, but not how to pin the deployment onto a selected node_group.
What I got is a single control plane, and then 3 node groups, 1. app, 2. db, 3.management.
I want to create the monitoring namespace and pin it onto the management node_group. and as this is EKS, PodNodeSelector is not a option,
due to size of the entire stack it's not practical to add a nodeSelector to each pod deployment yaml file... to many files to edit, to much chance for mistakes...
@@MarcelDempers ... thinking... I can do a complete directory search for : nodeSelector and then add the relevant selector to all the code, might a option... a nicer option would of course be if I could create a namespace and pit it to a node_group
@@georgelza It's often that folks want to patch a deployment without having to edit all the files.
Take a look at `kubectl patch -n namespace deployment mydeployment --patch $(cat patch.yaml)
This allows you to bundle a set of patches along with all your templates
In patch.yaml, you add only the node selector
spec:
template:
spec:
nodeSelector:
foo: bar
@@MarcelDempers my next ask is to get persistent volume added to the script. wanted to use EFS, but been informed Prometheus is not approved for EFS, has to be EBS.
Brilliant...
high quality all around. really enjoyable to watch. thanks for breaking the subject down to a newbie like myself
Excellent presentation skill. Thanks.
Hi Mercel, I'm following your step, but when I'm trying to download kubectl from google I'm getting *The specified key does not exist*, do you know if the url changed? Thanks. Awesome tutorial.
Just checking quickly, when you say use ekscli... and create the control plane... my VPC already have 2 subnets in 2 AZ's...
we know for the control plane, you actually really want it deployed across 3 AZ's. so question, as the control plane is a managed service, does it get create inside my VPC, (other words I need to have 3 subnets) or is it done outside my VPC, and all that goes into my VPC is the EC2 instance (which implies I can deploy them across my 2 configured AZ's)
I can't recall off the top of my head but as far as I'm aware, a managed cluster does not rely on your subnets as the control plane will run on Amazon's network and not your VPC.
@@MarcelDempers My thinking also, but seem to remember you providing subnet names and security group names when you created the management plane, let me look at your video again.
Keep up the good work !
Thanks for the video. Very helpful.
Blerry awesome man, gratitude
Excellent. Please post aws cicd for eks cluster
All of these would work for you 💪🏽czcams.com/play/PLHq1uqvAteVsSsrnZimHEf7NJ1MlRhQUj.html
Thanks for the video. I created role and attached policies for the nodes but, I get this error while creating the node group.
"An error occurred (InvalidRequestException) when calling the CreateNodegroup operation: Failed to create service linked role: AWSServiceRoleForAmazonEKSNodegroup due to missing permissions for 'iam:CreateServiceLinkedRole'
"
Hi Marcel, thanks for the video. I have a question, when we create our EKS cluster (through CLI or web console), does AWS create a separate VPC to host the control plane while attaching resources and ENI to our specified VPC and subnets? Thank you!
Hello! Fantastic tutorial and explanations!
I have a question: Where did you got the docker whale from ? :D I want one too so bad.
Thanks for the kind words 💪🏽
I happened to find it at a Chinese shop and it looks exactly like the Docker whale
😀
unable to find the docker-development-youtube-series/kubernetes/amazon path at repo. :S
github.com/marcel-dempers/docker-development-youtube-series/tree/master/kubernetes/cloud/amazon
Thank you!
Thanks for the video
Good one for beginners !!
Thanks ! Is a good practice to replace aws-cli with an terraform aproach?
I would not really consider either as best practise. They are both tools that solve different things. aws-cli helps you talk to AWS to provision resources.
It's a key to understand aws-cli when working with AWS.
Terraform helps provision infrastructure as code and can help manage infra at bigger scale, so it's also important to learn.
It's always best to understand aws-cli as an AWS user first, then take a look at Terraform to see what it solves for you.
Flex mate.
why is this guy so HUUUGEEEEE!!
i keep getting this error when defining worker node policy
aws iam create-role --role-name getting-started-eks-role-nodes --assume-role-policy-document .assume-node-policy.json
An error occurred (MalformedPolicyDocument) when calling the CreateRole operation: This policy contains invalid Json
Why does it sound like you're afrikaans.?
want ek kan die taal praat 🤣💪🏽👊🏽
@@MarcelDempers Jinne, dit is goeie nuus. As jy ooit in Midrand is moet ons 'n braai reel.