Welcome to the Linux Foundation Forum!

Does anyone else get frustrated by the amount of errors there are in the LFD259 training course?

I've noticed many errors in the text of the learning course and i find that really frustrating because i paid a lot of money for this course. An example of one such error:

In excercise 5.3 it says the following:

Review the YAML for our earlier simple pod. Recall that we added an Ambassador style logging container to the pod buthad not fully configured the logging

The use of the term ambassador is wrong here. It should be sidecar instead. Then below this text we have:

Let us begin by adding shared storage to each container. We will use the hostPath storage class to provide the PV and PVC. First we create the directory

hostPath is NOT a storage class. It is a storage type.

These are just two examples of terms used wrongly. I have encountered many such errors in the previous chapters as well. Someone from the linux foundation should really consider reading through the reading material and correct these errors. Also please do consider removing all the text that references flags or options that are scheduled to be removed in a future release. For instance the 'recycle' option for persistent volume API objects. Why would you want to teach such an option if it is going to be removed in a future release anyway?

Also, please update all the parts that mention using Docker as a container runtime inside a node. Docker can no longer be used for this in the more recent versions of kubernetes. Support for Docker has been removed years ago. So schematics used in the course should no longer use docker as a container runtime example.

A course people pay hundreds of dollars for should not contain these kinds of errors.

Answers

  • chrispokorni
    chrispokorni Posts: 2,585

    Hi @m.ramah,

    Thank you for identifying the content that needs reviewing. Your recommendations shall be passed on to the course maintainer.

    You may notice that very few features in the training material return deprecation warnings, but they are still worth mentioning hence they are incorporated in hands-on exercises. Deprecation is a gradual process that spans over a longer time period. Deprecated features frequently remain functional for months or even years, and in some cases, they are never fully removed. These deprecated features often still hold significant practical value and offer excellent learning opportunities for understanding core concepts.

    Regarding Docker, it still remains a fully viable container runtime for Kubernetes alongside options like containerd, cri-o, and Mirantis container runtime (MCR). What actually changed is the integration mechanism with Kubernetes, which now utilizes a shim externally maintained by Docker and Mirantis. I encourage reading more on the subject in the official Kubernetes documentation, to fully understand the transition and its effect on the Kubernetes container orchestration.

    Regards,
    -Chris

  • m.ramah
    m.ramah Posts: 8

    @chrispokorni i have another suggestion for improvement that you may communicate to the course author if you want.

    At the bottom part of page 4 of lab 6.5, the course text instructs the reader to add an ingress rule to the network policy which should allow all TCP ingress traffic coming from all other pods within the same namespace. Then the course instructs to run a test pod called test that does 5 ping requests to the ip address of another pod.

    Lastly it mentions that none of these ping requests should have gotten a response, with 100% packet loss. This is all correct but still confusing for the reader. The reason why it is confusing is because the course does not explain why the ping requests don't come through. Namely because the ingress allow rule is defined for the TCP protocol. But a Ping request uses the protocol ICMP instead. This may not be obvious for everyone.

  • m.ramah
    m.ramah Posts: 8

    Question 2 of Domain review 6.6 may also need to be changed or at least get some more explanation. The question is:

    Create a new LoadBalancer service to expose the newly created deployment. Test that it works.

    The problem with this is that the Loadbalancer type service has not yet been covered in the previous slides. Only ClusterIP and NodePort were explained.

  • chrispokorni
    chrispokorni Posts: 2,585
    edited February 26

    Hi @m.ramah,

    It is best to attempt the Domain review scenarios after the completion of the entire training course.

    Regards,
    -Chris

Categories

Upcoming Training