Don't miss out on new posts! Sign up! Also, visit the VHT Forums!
Kubernetes pods not...
 
Notifications
Clear all

Kubernetes pods not ready and internet connectivity problems after deploying calico

8 Posts
2 Users
1 Likes
222 Views
(@phantomsixthplayer)
Posts: 3
Active Member
Topic starter
 

After applying the tigera-operator.yaml and custom-resources.yaml I lose internet connectivity on k8s-master-1 and k8s-node-1. It seems when I sudo nmcli connection modify eth0 ipv4.dns "8.8.8.8" or add it to /etc/resolv.conf they show as READY.

The following pods calico-apiserver, csi-node-driver, csi-node-driver do not appear in a READY state. As this is my first Kubernetes deploy I noted the exact steps to configure on my GitHub as I may have missed a configuration or done something incorrectly.

I have also attached containerd and kubelet service logs after a vacuum.

 

hyper v manager network adapter

Any help much appreciated.

 
Posted : 04/01/2024 3:02 am
Brandon Lee reacted
Brandon Lee
(@brandon-lee)
Posts: 397
Member Admin
 

@phantomsixthplayer welcome to the forums? Did you mean to make the thread private? Let me know if we could make it public. It helps with visibility on the site. Thank you again, hopefully we can help you pinpoint the issue.

 
Posted : 04/01/2024 8:27 am
Brandon Lee
(@brandon-lee)
Posts: 397
Member Admin
 

@phantomsixthplayer can you post the contents of the tigera-operator.yaml and custom-resources.yaml?

 
Posted : 05/01/2024 7:11 am
Brandon Lee
(@brandon-lee)
Posts: 397
Member Admin
 

@phantomsixthplayer it looks like maybe your post reply didn't make it?

 
Posted : 05/01/2024 8:54 am
(@phantomsixthplayer)
Posts: 3
Active Member
Topic starter
 

I've uploaded it as log files instead. It seems permission issues with wordpress and user using codeblocks. I can quote the first post that was blank and it shows the code block I tried to post.

Posted by: @phantomsixthplayer

@brandon-lee 

hey mate, thanks for the reply. I believe its working now.
- With Hyper-V needed to not use the Default Switch and create a custom network adapter/nat as Hyper-V would change the IP
- Needed to assign static ip addresses to control-plane and worker-nodes and also add google dns
- Added the static entries to hosts file on the control-plane and worker nodes
- sudo kubeadm init --pod-network-cidr=192.168.0.0/16 was overlapping with home network probably wasn't good
- When applying the Calico custom-resource.yaml I needed to modify cidr: 192.168.0.0/16 to the one I decided for the sudo kubeadm init --pod-network-cidr= and then needed to apply this yaml i modified

I have some other errors in the kubelet.service and containerd.service. Not sure if they are okay to leave as is or need fixing.

 

 

 
Posted : 05/01/2024 2:05 pm
Brandon Lee
(@brandon-lee)
Posts: 397
Member Admin
 

@phantomsixthplayer Awesome good troubleshooting on that one! Glad you figured out the issue with your Hyper-V switches. Are you finding Hyper-V easy to work with running Kubernetes nodes? I am wondering how many are running Hyper-V in their labs actually. That number may be going up with the VMware buyout.

 
Posted : 05/01/2024 10:06 pm
(@phantomsixthplayer)
Posts: 3
Active Member
Topic starter
 

Posted by: @brandon-lee

@phantomsixthplayer Awesome good troubleshooting on that one! Glad you figured out the issue with your Hyper-V switches. Are you finding Hyper-V easy to work with running Kubernetes nodes? I am wondering how many are running Hyper-V in their labs actually. That number may be going up with the VMware buyout.

 

So far running it on Hyper-V its going great. Only thing I dislike about Hyper-V Manager on Windows 10 is the export or clone in other virtualisation software, though I haven't invested time into it yet.

Its my first time setting up Kubernetes and boy that was a challenge. I do still have some kubelet.service and containerd.service odd messages which I can't work out. I've attached them to see if anyone has any ideas.

 

 

This post was modified 2 months ago by phantomsixthplayer
 
Posted : 06/01/2024 5:32 am
Brandon Lee
(@brandon-lee)
Posts: 397
Member Admin
 

@phantomsixthplayer, checking back in. Were you able to get things running smoothly? Also, have you overcome the errors you were seeing in the logs.

 
Posted : 18/01/2024 9:49 am