Given a container that writes a log file in format A and a container that converts log files from format A to format B, create a deployment that runs both containers such that the log files from the…

Question

while true; do
 echo "i luv cncf" >> /tmp/log/input.log;
 sleep 10;
done

i luv cncf
i luv cncf
i luv cncf

Exhibits

Explanation

The task requires creating a Kubernetes Deployment with two containers, sharing an emptyDir volume, and mounting a ConfigMap to the sidecar, demonstrating proficiency in multi-container pod design and kubectl YAML management.

Approach. To correctly complete the task, the user must perform the following interactions in the terminal:

1. Create the ConfigMap: The task specifies that the ConfigMap for Fluentd is provided at /opt/KDMC00102/fluentd-configmap.yaml. This ConfigMap must be created first so the Deployment can reference it.

   kubectl create -f /opt/KDMC00102/fluentd-configmap.yaml
   

   Reasoning: The adapter-zen sidecar container needs this ConfigMap mounted to /fluentd/etc. The Deployment will fail to create the Pod if the referenced ConfigMap does not exist.

2. Create/Edit Deployment YAML: The core of the task is to define a Kubernetes Deployment. The user should create a new YAML file (e.g., deployment-xyz.yaml) with the following content (or generate a basic deployment and modify it using vim):

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: deployment-xyz
     labels:
       app: logger-adapter # Label for selector
   spec:
     replicas: 1
     selector:
       matchLabels:
         app: logger-adapter
     template:
       metadata:
         labels:
           app: logger-adapter
       spec:
         # Define volumes for the pod
         volumes:
         - name: log-volume
           emptyDir:
             medium: Memory # Use tmpfs for non-persistent, in-memory storage
         - name: fluentd-config
           configMap:
             name: fluentd-configmap # Assuming the ConfigMap file defines this name
         # Define containers for the pod
         containers:
         - name: logger-dev
           image: lfccncf/busybox:1
           command: ["sh", "-c"]
           args: # Use args for multi-word commands
           - while true; do echo "i luv cncf" >> /tmp/log/input.log; sleep 10; done
           volumeMounts:
           - name: log-volume
             mountPath: /tmp/log # Mount shared log volume
         - name: adapter-zen
           image: lfccncf/fluentd:v0.12
           volumeMounts:
           - name: log-volume
             mountPath: /tmp/log # Mount shared log volume
           - name: fluentd-config
             mountPath: /fluentd/etc # Mount ConfigMap for Fluentd config
   

   Reasoning: This YAML file correctly defines a Deployment named deployment-xyz with one replica, containing a pod with two specified containers (logger-dev and adapter-zen). It configures an emptyDir volume for shared logging and a configMap volume to inject Fluentd configuration, as required by the task. The command and args are specified correctly for the logger-dev container.

3. Apply the Deployment: Once the deployment-xyz.yaml file is created and saved, apply it to the Kubernetes cluster.

   kubectl apply -f deployment-xyz.yaml
   

   Reasoning: This command creates or updates the Deployment resource based on the provided YAML, which will then manage the creation of the Pod with the two containers.

Common mistakes.

• common_mistake. A common mistake would be to create a standalone Pod resource instead of a Deployment. The question explicitly asks for a 'deployment named deployment-xyz'. Creating a Pod directly would fulfill the immediate container requirements but would lack the self-healing, scaling, and update capabilities that a Deployment provides. Another mistake, hinted at by the exhibit, is incorrect YAML syntax or structure, such as misplacing mountPath (it should be nested under volumeMounts within a container definition) or incorrect indentation. Forgetting to create the ConfigMap before the Deployment, or misspelling its name in the Deployment YAML, would also lead to an error. Incorrectly defining the emptyDir volume or failing to mount it to both containers, or omitting the ConfigMap mount for the adapter-zen sidecar, would result in the log conversion not functioning as intended. Misconfiguring the command and args for the logger-dev container would prevent it from writing logs.

Concept tested. This question primarily tests the understanding and application of Kubernetes Deployments, multi-container Pod design (sidecar pattern), shared volumes (emptyDir), ConfigMaps for configuration injection, and kubectl CLI proficiency for resource creation and YAML manipulation.

No community discussion yet for this question.