Kubernetes: How to Create a deployment
In this article, I will show you how to create a straightforward Kubernetes deployment! to make things simple; the deployment will be…
In this article, I will show you how to create a straightforward Kubernetes deployment! to make things simple; the deployment will be something straightforward, a pod that will ping 1.1.1.1 using the alpine image, then we will perform various operations like scaling, getting logs, deleting pods and the deployment. I assume that you already have a Kubernetes deployment.
What is a deployment?
A Kubernetes deployment is an object in the Kubernetes cluster that describes how to deploy an application. It defines the desired state of the application, such as the number of replicas, the container image to use, and the container ports to expose. The deployment object also manages the rollout and rollback of the application, ensuring that the desired state is maintained.
Create a deployment
To create a deployment with a single pod that will ping 1.1.1.1 enter the following
$ kubectl create deployment pingpong --image alpine -- ping 1.1.1.1
deployment.apps/pingpong createdGet the status of the deployment and pods
To get the status of the deployment, we use the “get deployment” command
$ kubectl get deployments
NAME READY UP-TO-DATE AVAILABLE AGE
pingpong 1/1 1 1 47sThis means that this deployment has one pod and is ready; we can verify this on the pod level using the “get pods” command
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
pingpong-6d8f458c56-f2bgl 1/1 Running 0 2m29sScale up the deployment
To scale up the deployment to three pods running in parallel, we use the “scale deployment” command
$ kubectl scale deployment/pingpong --replicas 3We can verify that our deployment now has three pods!
$ kubectl get deployments
NAME READY UP-TO-DATE AVAILABLE AGE
pingpong 3/3 3 3 9m23s$ kubectl get pods
NAME READY STATUS RESTARTS AGE
pingpong-6d8f458c56-9x47c 1/1 Running 0 91s
pingpong-6d8f458c56-bkk9r 1/1 Running 0 91s
pingpong-6d8f458c56-f2bgl 1/1 Running 0 10mWhat happen if i delete a pod of the deployment?
What do you believe will happen if I delete a pod from the deployment? let’s see
$ kubectl delete pod pingpong-6d8f458c56-9x47c
pod "pingpong-6d8f458c56-9x47c" deletedNow let’s get the status of the pods again
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
pingpong-6d8f458c56-9x47c 1/1 Terminating 0 3m
pingpong-6d8f458c56-bkk9r 1/1 Running 0 3m
pingpong-6d8f458c56-f2bgl 1/1 Running 0 12m
pingpong-6d8f458c56-rcsv2 1/1 Running 0 8sWe can see that the pod we are deleting is in the terminating phase, and another one has been created in the last 8seconds; this is happening because we have instructed Kubernetes, no matter what, always to have a deployment with three pods.
How to get logs from pods and deployments
To view the logs from a single pod, we need to know its name and then use the logs <pod_name> -f tail -1 command; if you are familiar with Linux, you probably know that the tail -f commands prints the log since the last line and then after, -f tail -1 achieves the same functionality
$ kubectl get pods
NAME READY STATUS RESTARTS AGE
pingpong-6d8f458c56-bkk9r 1/1 Running 0 7m33s
pingpong-6d8f458c56-f2bgl 1/1 Running 0 16m
pingpong-6d8f458c56-rcsv2 1/1 Running 0 4m41s
kpatronas@nautilus:~$ kubectl logs pingpong-6d8f458c56-bkk9r -f --tail 1
64 bytes from 1.1.1.1: seq=454 ttl=37 time=9.490 ms
64 bytes from 1.1.1.1: seq=455 ttl=37 time=9.224 ms
64 bytes from 1.1.1.1: seq=456 ttl=37 time=9.344 ms
64 bytes from 1.1.1.1: seq=457 ttl=37 time=11.733 msIn case you want the full log, you can enter logs <pod_name>; thiswill print until the last line and will exit, it will not wait for new log lines to be appended
kubectl logs pingpong-6d8f458c56-bkk9r
64 bytes from 1.1.1.1: seq=627 ttl=37 time=10.917 ms
64 bytes from 1.1.1.1: seq=628 ttl=37 time=17.263 ms
64 bytes from 1.1.1.1: seq=629 ttl=37 time=9.081 ms
64 bytes from 1.1.1.1: seq=630 ttl=37 time=14.082 ms
64 bytes from 1.1.1.1: seq=631 ttl=37 time=13.970 ms
*
*
*
64 bytes from 1.1.1.1: seq=641 ttl=37 time=13.032 ms
$To get the logs from deployment is somehow misleading, you dont get logs from all the pods but just from some of the pods of the deployment
$ kubectl logs -f deployment/pingpong --tail 1
Found 3 pods, using pod/pingpong-6d8f458c56-f2bgl
64 bytes from 1.1.1.1: seq=1406 ttl=37 time=10.273 ms
64 bytes from 1.1.1.1: seq=1407 ttl=37 time=11.746 ms
64 bytes from 1.1.1.1: seq=1408 ttl=37 time=9.489 ms
64 bytes from 1.1.1.1: seq=1409 ttl=37 time=9.582 ms
64 bytes from 1.1.1.1: seq=1410 ttl=37 time=10.793 ms
64 bytes from 1.1.1.1: seq=1411 ttl=37 time=11.257 msHow to delete a deployment
To delete a deployment we need to use the “delete deployment” command
$ kubectl delete deployment/pingpong
deployment.apps "pingpong" deletedAfter a while if you check for deployments and pods you should not see any!
$ kubectl get deployment/pingpong
Error from server (NotFound): deployments.apps "pingpong" not found
$ kubectl get pods
No resources found in default namespace.Conclusion
In this article, we saw how to make a simple deployment, how to scale it, and how to manage it; of course, there are many things to cover, and the production way to do things is using YAML files, but I think for tutorial purposes not using YAML can make things much easier to understand.
