Welcome to the Linux Foundation Forum!

2.8 cant not pass, it seems like a network issue

Options



Both under the same network. Any advises? I am all ears! Thank you!

Answers

  • chrispokorni
    chrispokorni Posts: 2,224
    Options

    Hi @htian,

    Please review the setup video for GCE VMs, that is found in the introductory chapter. Aside from the same VPC and subnet, the firewall needs to allow all inbound traffic to the nodes.

    Regards,
    -Chris

  • htian
    htian Posts: 3
    Options

    @chrispokorni Thank your for your response. I have done and checked many times to made sure everything is followed.
    The VPC network has a firewall rule as in the video, please see below
    Now both nodes appear under same network (lfclass in this case)
    However, they were not able to ping/traceroute each other

  • chrispokorni
    chrispokorni Posts: 2,224
    Options

    Hi @htian,

    What are the outputs of:

    kubectl get nodes -o wide
    kubectl get pods -A -o wide

    I suspect you may have missed this note in the Overview section of lab exercise 2.1:

    Don’t use 10.0.0.0/8 network for nodes

    The 10.0.0.0/8 network is expected to be used by the Cilium network plugin for the Pod network. Having VMs/Nodes and Pods on overlapping networks will cause routing issues between nodes and between pods. A simple change in VM/Node network, such as 10.200.0.0/16 may be sufficient to allow you to continue with the exercise.

    Regards,
    -Chris

Categories

Upcoming Training