I've set up two Kubernetes cluster using kubeadm on both Azure and OpenStack, running on Ubuntu 24.04 with Kubernetes version 1.28. I'm using Calico as the CNI (Container Network Interface). However, I'm experiencing communication issues between my pods when trying to use curl or ping.
When the two pods that need to communicate are on different worker nodes, there's no response. But when the two pods are on the same worker node, the communication works fine.
I've opened all firewalls on both Azure and OpenStack, but that didn't change anything. I also tried using a different CNI (Flannel), but the issue persists.
Some informations about the cluster:
[control plane and worker nodes] https://k8s-sigl-2025.s3.eu-west-3.amazonaws.com/Capture+d%E2%80%99e%CC%81cran+2024-08-13+a%CC%80+00.03.52.png
[kube system namespace] https://k8s-sigl-2025.s3.eu-west-3.amazonaws.com/Capture+d%E2%80%99e%CC%81cran+2024-08-13+a%CC%80+00.01.53.png
[default namespace] https://k8s-sigl-2025.s3.eu-west-3.amazonaws.com/Capture+d%E2%80%99e%CC%81cran+2024-08-13+a%CC%80+00.06.44.png
When I analyze the network traffic on worker2 (192.168.0.4), I can see the ping request, but on worker1 (192.168.0.5), there's nothing showing up in tcpdump and so no response: https://k8s-sigl-2025.s3.eu-west-3.amazonaws.com/Capture+d%E2%80%99e%CC%81cran+2024-08-13+a%CC%80+00.00.40.png
Thanks in advance for your help; I've been stuck on this issue for several weeks now.
