{"id":6297,"date":"2019-07-26T18:57:43","date_gmt":"2019-07-26T18:57:43","guid":{"rendered":"https:\/\/www.devopsschool.com\/blog\/?p=6297"},"modified":"2021-11-16T11:11:38","modified_gmt":"2021-11-16T11:11:38","slug":"working-with-kubernetes-cluster-using-kubectl-part-10-pods-life-cycle","status":"publish","type":"post","link":"https:\/\/www.devopsschool.com\/blog\/working-with-kubernetes-cluster-using-kubectl-part-10-pods-life-cycle\/","title":{"rendered":"Working with Kubernetes Cluster using Kubectl Part – 10 – Pods Life Cycle"},"content":{"rendered":"\n

pod-restart-policy.yaml<\/strong><\/p>\n\n\n

apiVersion: v1\nkind: Pod\nmetadata:\n  name: hello-world-onfailure-pod\nspec:\n  containers:\n  - name: hello-world\n    image: gcr.io\/google-samples\/hello-app:1.0\n  restartPolicy: OnFailure\n---\napiVersion: v1\nkind: Pod\nmetadata:\n  name: hello-world-never-pod\nspec:\n  containers:\n  - name: hello-world\n    image: gcr.io\/google-samples\/hello-app:1.0\n  restartPolicy: Never\n<\/code><\/span><\/pre>\n\n
\n\n#Start up kubectl get events --watch and background it.<\/span>\nkubectl get events --watch &\nclear\n\n#Create a pod...we can see the scheduling, container pulling and container starting.<\/span>\nkubectl apply -f pod.yaml\n\n#We've used exec to launch a shell before, but we can use it to launch ANY program inside a container.<\/span>\n#Let's use killall to kill the hello-app process inside our container<\/span>\nkubectl exec -it hello-world-pod -- \/bin\/sh \nps\nexit<\/span>\n\n#We still have our kubectl get events running in the background, so we see if re-create the container automatically.<\/span>\nkubectl exec -it hello-world-pod -- \/usr\/bin\/killall hello-app\n\n#Our restart count increased by 1 after the container needed to be restarted.<\/span>\nkubectl get pods\n\n#Look at Containers->State, Last State, Reason, Exit Code, Restart Count and Events<\/span>\n#This is because the container restart policy is Always by default<\/span>\nkubectl describe pod hello-world-pod\n\n#Cleanup time<\/span>\nkubectl delete pod hello-world-pod\n\n#Kill our watch<\/span>\nfg\nctrl+c\n\n#Remember...we can ask the API server what it knows about an object, in this case our restartPolicy<\/span>\nkubectl explain pods.spec.restartPolicy\n\n#Create our pods with the restart policy<\/span>\nmore pod-restart-policy.yaml\nkubectl apply -f pod-restart-policy.yaml\n\n#Check to ensure both pods are up and running, we can see the restarts is 0<\/span>\nkubectl get pods \n\n#Let's kill our apps in both our pods and see how the container restart policy reacts<\/span>\nkubectl exec -it hello-world-never-pod -- \/usr\/bin\/killall hello-app\nkubectl get pods\n\n#Review container state, reason, exit code, ready and contitions Ready, ContainerReady<\/span>\nkubectl describe pod hello-world-never-pod\n\n#let's use killall to terminate the process inside our container. <\/span>\nkubectl exec -it hello-world-onfailure-pod -- \/usr\/bin\/killall hello-app\n\n#We'll see 1 restart on the pod with the OnFailure restart policy.<\/span>\nkubectl get pods \n\n#Let's kill our app again, with the same signal.<\/span>\nkubectl exec -it hello-world-onfailure-pod -- \/usr\/bin\/killall hello-app\n\n#Check its status, which is now Error too...why? The backoff.<\/span>\nkubectl get pods \n\n#Let's check the events, we hit the backoff loop. 10 second wait. Then it will restart.<\/span>\n#Also check out State and Last State.<\/span>\nkubectl describe pod hello-world-onfailure-pod \n\n#Check its status, should be Running...after the Backoff timer expires.<\/span>\nkubectl get pods \n\n#Now let's look at our Pod statuses<\/span>\nkubectl delete pod hello-world-never-pod\nkubectl delete pod hello-world-onfailure-pod\n<\/code><\/span>Code language:<\/span> PHP<\/span> (<\/span>php<\/span>)<\/span><\/small><\/pre>\n\n