Unterschiede

Hier werden die Unterschiede zwischen zwei Versionen angezeigt.

--- en:modul:m321_aws:topics:08 [2025/09/27 22:58] – dgaravaldi
+++ en:modul:m321_aws:topics:08 [2025/10/22 16:32] (aktuell) – dgaravaldi
@@ Zeile 22: / Zeile 22: @@
 \\
-===== Example 1: Kubernetes Deployment (YAML) =====
+==== Example 1: Deployment-Elements in YAML explained ====
+<code>
+apiVersion: apps/v1
+kind: Deployment                 (1)
+metadata:
+  name: calculator-deployment    (2)
+  labels:
+    app: calculator
+spec:
+  replicas: 3                    (3)
+  selector:                      (4)
+    matchLabels:
+      app: calculator
+  template:                      (5)
+    metadata:
+      labels:                    (6)
+        app: calculator
+    spec:
+      containers:
+      - name: calculator         (7)
+        image: gcr.io/calculator (8)
+        ports:                   (9)
+        - containerPort: 8080
+</code>
+In this YAML configuration, we have to ensure the following:
+\\
+  -We have defined a Kubernetes resource of the Deployment type from the ''apps/v1'' Kubernetes API version.
+  -The unique deployment name is ''calculator-deployment''.
+  -We have defined that there should be exactly ''3'' of the same Pods created.
+  -''selector'' defines how Deployment finds Pods to manage, in this case, just by the label.
+  -''template'' defines the specification for each created Pod.
+  -Each Pod is labeled with ''app: calculator''.
+  -Each Pod contains a Docker container ''named calculator''.
+  -A Docker container was created from the image called ''gcr.io/calculator''.
+  -The Pod exposes container port ''8080''.
+To install the deployment, run the following command:
+<code>
+$ kubectl apply -f deployment.yaml
+</code>
+You can check that the three Pods, each containing one Docker container, have been created:
+<code>
+$ kubectl get pods
+NAME                                  READY STATUS  RESTARTS AGE
+calculator-deployment-dccdf8756-h2l6c 1/1   Running 0        1m
+calculator-deployment-dccdf8756-tgw48 1/1   Running 0        1m
+calculator-deployment-dccdf8756-vtwjz 1/1   Running 0        1m
+</code>
+Each Pod runs a Docker container. We can check its logs by using the following command:
+<code>
+$ kubectl logs pods/calculator-deployment-dccdf8756-h2l6c
+</code>
+\\
+==== Example 2: Kubernetes Deployment (YAML) ====
 As an example, containerized python app is used. The parameter ''terminationGracePeriodSeconds'' is set to give your app enough time to finish work.
@@ Zeile 48: / Zeile 109: @@
 </code>
 \\
-===== Example 2: Kubernetes Deployment + Service (YAML) =====
+==== Example 3: Kubernetes Deployment + Service (YAML) ====
 The previous example declared only the deployment. In this example, we add the corresponding service.
@@ Zeile 84: / Zeile 145: @@
       targetPort: 8080
   type: ClusterIP
+</code>
+\\
+\\
+===== Service types =====
+As a recapture on how your application can be accessed from the outside, we need to start with the types of Kubernetes Services. You can use four different service types, as follows:
+  * ''ClusterIP (default)'': The service has an internal IP only.
+  * ''NodePort'': Exposes the service on the same port of each cluster node. In other words, each physical machine (which is a Kubernetes node) opens a port that is forwarded to the service. Then, you can access it by using ''<NODE-IP>:<NODE-PORT>''.
+  * ''LoadBalancer'': Creates an external load balancer and assigns a separate external IP for the service. Your Kubernetes cluster must support external load balancers, which works fine in the case of cloud platforms, but may not work if you use minikube.
+  * ''ExternalName'': Exposes the service using a DNS name (specified by externalName in the spec).
+While LoadBalancer seems to be the simplest solution, it has two drawbacks:
+  * It's not always available, for example, if you're using minikube.
+  * External public IPs are usually expensive. A different solution is to use a ''NodePort'' service.
+<code>
+$ kubectl get service my-python-service
+NAME               TYPE     CLUSTER-IP    EXTERNAL-IP  PORT(S)        AGE
+my-python-service  NodePort 10.19.248.154 <none>       8080:32259/TCP 13m
+</code>
+You can see that port ''32259'' was selected as a node port. This means that we can access our ''my-python-service'' service using that port and the IP of any of the Kubernetes nodes.
+The IP address of your Kubernetes node depends on your installation. In the case of minikube, you can check it with the minikube ip command:
+<code>
+$ kubectl get nodes -o jsonpath='{ $.items[*].status.addresses[?(@.type=="ExternalIP")].address }'
+.192.180.252 35.232.125.195 104.198.131.248
+</code>
+To check that you can access ''my-python-service'' from the outside, run the following command:
+<code>
+$ curl http://<NODE-IP>:32047/
 </code>