Welcome to the Linux Foundation Forum!

Problems creating Daemon

Hello,

I am trying to reproduce the lab "Demo: Managing Node Applications with DaemonSets" on my local environment.

I have reproduced the daeomon yaml file as follows:

but then, despite following the commands from the video. Result was not the expected (while I still had the deployment with 3 nodes from the former demo running).

When checking kubernetes dashboard, I could see the following error.

Please kindly let me know if there is something wrong in the yaml template I created to reproduce the daemon yaml file locally, or kindly suggest me what I could be doing wrong.

Thanks in advance for your time and help.

Sincerely,

Josep Maria

Best Answer

  • chrispokorni
    chrispokorni Posts: 2,372
    Answer ✓

    Hi @josepmaria,

    Perhaps there is a demo video you may have missed where multi-node clusters are discussed.
    However, you can scale a cluster with the minikube node add command. To find out more about its usage, please visit the minikube command reference.

    Regards,
    -Chris

Answers

  • chrispokorni
    chrispokorni Posts: 2,372

    Hi @josepmaria,

    You can simply replace the image with another OCI compliant image, such as fluentd hosted on docker hub. The updated manifest should look like this:

    ...
        spec:
          containers:
          - name: fluentd-agent
            image: fluentd
          terminationGracePeriodSeconds: 30
    

    Regards,
    -Chris

  • josepmaria
    josepmaria Posts: 80
    edited October 8

    Hi @chrispokorni ,

    Thank you for your time and suggested solution.

    Unfortunately, it does not work either.

    Kindly see updated yaml file and the deamonset successful creation:

    Unfortunately, despite having the deployment from former demo running, newly created daemonset does not install in all nodes.

    Any advise will be highly appreciated. Thanks in advance for your time and support.

    Sincerley,

    Josep Maria

  • chrispokorni
    chrispokorni Posts: 2,372

    Hi @josepmaria,

    Depending on the size of your cluster this may be the expected outcome. Keep in mind that the DaemonSet Pods are also subject to scheduling constraints such as taints/tolerations, namespaces, etc...

    What are the outputs of:

    kubectl get nodes -o wide
    kubectl describe nodes | grep -i taint
    

    Regards,
    -Chris

  • josepmaria
    josepmaria Posts: 80
    edited October 8

    Hello @chrispokorni ,

    Thank you for your time and patience, I appreciate your support.

    Here is a screenshot with the commands and output (newly requested commands are at the end of the picture below):

    Looking forward to hearing from you, I thank you in advance for you feedback.

    Sincerley,

    Josep Maria

  • chrispokorni
    chrispokorni Posts: 2,372

    Hi @josepmaria,

    Based on the provided output, you have one node, hence the only one DaemonSet Pod created. Did you expect a different outcome? Why?

    For a better understanding of the DaemonSet controller and its use case, please revisit the course content and the official documentation for additional context:
    https://kubernetes.io/docs/concepts/workloads/controllers/daemonset/

    Regards,
    -Chris

  • josepmaria
    josepmaria Posts: 80

    Hi @chrispokorni ,

    Thanks for your quick answer, I appreciate it.

    Yes, I expected a different outcome. I have been doing my best to replicate the code used in the labs into my local host minikube.

    On the previous demo, I had deployed the appliation and scaled it up to 3 replicasets (just like in the video).

    Since in your demo, you had 3 nodes, I thought I would also have 3 nodes.

    I understand daemonset is installed in each node, thus if there is only one node, then my installation is fine.

    Allow me to ask you, was there any node increase in the former video via the command line? Perhaps I missed it and thus ended up using only one node.

    Once again, thanks for your time and support.

    Sincerely,

    Josep Maria

  • Hi @chrispokorni ,

    Thank you.

    You are right, I had not correctly implemented the multi-node cluster as per your explanation the the Lesson 7 demo : Getting Started with Minikube and Profiles.

    I have now deleted my current cluster and created a new one as per the video specification, I installed Kubctl, created and run the daemon.

    As a result, I finally get the expected output:

    Once again, thank you for your time and patience. I really appreciate it.

    Sincerely,

    Josep Maria

Categories

Upcoming Training