Could you make more Rancher videos? it's a popular topic and there isn't much tutorials about making a self hosted Rancher setup with low end hardware specs ( load balancing )
Great video, thanks. It helped me in my project. i have already installed k3s on my two servers and up and running in less than the time I used to use installing k8s with kubeadm
@@CloudVersityOfficial Not harder than setting up a cluster for the first time, while learning how to do it, and all this for production environment:)) One of the super useful things you showed was the nip.io service - I didn't know about it. However, it doesn't seem to fit a production environment approach, so it would be awesome if you could explain like what are the most basic ways to expose services like a database to a local network in which the k3s cluster is located, and a service to the internet. In my case, I already exposed those using simple nodeport, but I know it's not ideal.
@@redmictian Thanks for the input. I'll put that on my todo list. :) Routing and securely exposing stuff always causes headaches and gets complicated pretty fast but I'll see what I can do. nip.io is for playground environments, as soon as you go in the direction of production environments your goal should be to have a valid DNS.
Thank you very much for your video. I'm doing a k3s cluster across 3 vm with tailscale and thanks to your tutorial I was able to create the server & node easly. I just got an error deploying the manifest ..... dial tcp cluster-ip:443 io/timeout ..... and if I find the solution I'll reply here.
Well I couldn't find a solution, I tried again reinstalling my vps and using your cmd but I end up always with the same error msg: Error from server (InternalError): error when creating "manifest.yaml": Internal error occurred: failed calling webhook "validate.nginx.ingress.kubernetes.io": Post "ingress-nginx-controller-admission.ingress-nginx.svc:443/networking/v1beta1/ingresses?timeout=10s": dial tcp 10.43.35.60:443: connect: connection refused
@@ugurserhattoy same issue: just delete the ingress-nginx-controller-admission in ingress-nginx namespace or just reinstall ingress-nginx by helm method. That's it.
i'm trying to install k3s on a Raspberry Pi 4 Model B 4Gb, but i have some problem. First problem, when i install the master node, the installation give me a this error "Job for k3s.service failed because the control process exited with error code. See "systemctl status k3s.service" and "journalctl -xeu k3s.service" for details." Second problem, when i use " cat /var/lib/rancher/k3s/server/node-token" the console give me the error "permission denied". I don't know where is the problem :(
Is the process in the video only meant to work with linode VPS's, or should it work with google cloud VPS's as well? I I'm not sure why, but the process is not working for me with VPS's on google cloud. After installing ingress-nginx, "kubectl -n ingress-nginx get svc" is showing internal IP-addresses on the column "External IP" instead of the external ones. The VPS's have the port 80, 443, 6443 already open though. Did anybody have a similar problem? I checked the yaml file for installing ingress-nginx, and I couldn't find anything obvious.
Good question :) I ran into similar problems with AWS, Azure and GCP, that's why I like to use simple cloud providers like Linode for these kind of demos because they are easier to handle. The big cloud providers will give you a fake public ip address which seems to not be know by your actual VM. Just check it with "netstat -i" or similar. They just get NATed down the road until it hits your VM in their "private" network. This is just an assumption, I'm not 100% sure if this is a correct explanation so take it with caution. However there is a fix. You need to manually edit the ingress nginx svc "ingress-nginx-controller" and manually add spec.externalIPs: [] your public IP. A "kubectl -n ingress-nginx get svc" should now update and also list your public IP besides the private one. You can now proceed with the demo. I also updated the the ingress by adding the "spec.ingressClassName" which is necessary since v1.22. Cheerio.
@@CloudVersityOfficial Sorry a bit late, but.. I just tried what you suggested, and worked great! Thanks. This "problem" made me learn even more about kubernetes :) FYI, it looks like I needed to edit the services on both "ingress-nginx-controller" and "ingress-nginx-controller-admission", and add the external IPs on both. Cheers.
Good video. Is there a reason why not keep the default ingress controller (Treafik)? Also, would you give thoughts on how to approach a two-node cluster (e.g. like to Pis) where one is a redundant of the other in case the first fails. No chance to have 3rd or more.
Hej dude, I'm just more familiar with Nginx. If you like Treafik go with it ;) K8s/K3s stands for high availability, so if one node fails it'll always try its best to keep all pods running on another node. Of course you need to make sure, that your "brain" the master node is deployed in a HA setup as well, otherwise your worker nodes would be kind of "brainless". There shouldn't be a redundant server, all of them can work and balance their load in terms of deployed pods and resource management.
I just recognized, that I linked to the wrong Twitter profile in this video :D 🔥 Social Media 🔥 Twitter ► twitter.com/cloud_versity Instagram ► instagram.com/cloudversity
15:37 - Hi, how do you have the public IP's for EXTERNAL-IP? I get the private IPs instead. And also can you update this video with the newest versions? Thank you for your efforts and time
I'm glad that it was helpful to you ;) Actually k3s will deploy workloads per default also on the server nodes. If you want to have you control plane separated from your workloads you can leverage from "taints". Here is a link to the k3s documentation which covers that point: rancher.com/docs/k3s/latest/en/installation/ha/#2-launch-server-nodes You can just add the option to the install command like: curl -sfL get.k3s.io | sh -s - server --node-taint CriticalAddonsOnly=true:NoExecute This should do the trick :)
Nope, just installing k3s and any sort of ingress controller won't give you ssl certificates out of the box. Feel free to check out a video I made exactly about this topic :) ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-deLW2h1RGz0.html