Ubuntu 18.04 issue

[Nahum48]

Hi,

As you can see in my screen shot, i'm having trouble with working on the K3S cluster. i followed the installation instructions using Jeff geerling Youtube channel.

My environment is working with VM's not RPI. I have 3 workers and 1 master.

when i'm trying to browse to the :

prometheus-k8s prometheus.192.168.1.39.nip.io 192.168.1.37 80, 443 89m alertmanager-main alertmanager.192.168.1.39.nip.io 192.168.1.37 80, 443 89m grafana grafana.192.168.1.39.nip.io 192.168.1.37 80, 443 89m

all failed, and you can see an example with the screen shot i attached.

anyone know if there's an issue on running the cluster on Ubuntu?

[magikmw]

Your nip.io addresses point to 192.168.1.39, but your IP addresses on ingress is 192.168.1.37. Check the masterIP and suffixDomain vars in vars.jsonnet.

[Nahum48]

Hi Magi and thanks for the comment.

2 questions please.

1. does the addresses have to be the same in the mip.io and the ingress.io?
2. I'm working in a home lab, not an organization domain, what's the DNS that i need to use? (does it necessary)

[magikmw]

1. If you're using the nip.io domain the IP number in them has to align with the IP adress of the ingress. (so both columns HOST and ADDRESS have to align in kubectl get ingress -n monitoring output.
2. nip.io is a special domain setup in a way that makes it feasible to use in this case - it points to any IP adress you'll put in as subdomain for *.nip.io. It's actually perfect to use if you don't have your own domain and DNS setup. Read more here: https://nip.io/

[Nahum48]

Thank you magikmw for your quick reply.

I'll read more about it, and apply the settings in my Master host machine.

[Nahum48]

my journey to make this cluster working....

new issue today, now instead of the error message that the page can not be displayed when i;m getting an error "service is not available" and as you can see in the screen shot i added, i have ALOT of CrashLoopBackOff so what is that, and how can i debug it.

As always, thank you for your answer, and remember i'm a rookie in all this Kubernetes world, so bear with me. :)

p.s i also added a screen shot of my vars.jsonnet, (maybe it'll help although i settings the settings are correct)

[magikmw]

Your nip.io and IP is still misaligned. It seems like your ingress controller uses 192.168.1.17, but you're configuring cluster-monitoring to use 192.168.1.37 (it's in output of kubectl get ingress -n monitoring). Can't help you much more without getting into your cluster networking (maybe IP adresses change due to DHCP?).

As for the CrashLoopBackOff - try using kubectl describe command to see events generated while the pods were created. This article helped me a bunch before LINK.

[Nahum48]

Hi there,

Good news, seems like we are in the right way to success. as you suggested, i changed the IP address to 192.168.1.17 in the vars file. and things started working. :+1:

As for the errors, can this give you any point what's the issue with the routing? my firewall service is down. I don't have IP Tables installed

[magikmw]

I thought you need iptables for the internal networking magic? As for the errors, not sure. It seems like there's trouble reaching the metrics service. Maybe it's not installed or there's trouble with networking.

[Nahum48]

Hi Magi,

I'll get back on the cluster on Sunday and see if there are networking issue (communications between the VM's) and i don't use IP Tables (it's not installed on my VM).

[carlosedp]

This appears to be the host configuration and not the monitoring stack issue.

Closing.