DevOps Classroomnotes 09/Sep/2023

Pods Contd

• Scaling: Scaling in k8s means increasing number of Pods not containers in Pod. For Scaling pods we would learn Replica set/Replication Controller etcs..

How k8s identifies objects: Labels

• Label is a key pair examples are
  • app: nginx
  • version: v1.0
• Labels are used to select/query kubernetes objects
• Refer Here for official docs on labels
• Selectors in k8s help in querying objects using labels
• selectors are of two types
  • Equality Based Selectors
  • Set based selectors

Pod with Labels

• Overview:
  
• Refer Here for the changes done
• Create pods
  

Pod with container with additional commands

• I have to add a container to the pod in addition to nginx which has a docker command docker container run -d alpine sleep 1d
• Refer Here for the changes
  

Interacting with containers

• In docker docker container exec -it or docker container exec
• In k8s we have kubectl exec
  
• Interactive tty
  

What happens when container goes into exited state

• Lets create a pod with alpine container which exits
• Refer Here for the changes
• kubernetes tries to restart containers, if it fails continuous k8s keeps trying and gives the status as CrashLoopBackoff
  
• Kuberentes restarts the containers always as the restartPolicy Default value is Always, which we can change Refer Here

Container types in Pod

• init containers: These containers are created prior to actual/main containers. ideally these containers should be short lived and majorly for meeting preconditions to run your application. Refer Here for official docs
• containers: This is where we run actual applications and they are expected to be living forever (continously)

Demonstration:

• Refer Here for the spec
  

Pod Lifecycle

• Refer Here for the Pod lifecycle
  

Activity (in docker)

• Create a nginx container with 128 MB of RAM
• Create a jenkins continer with “0.5” cpu and 256 MB of RAM

docker container run  -P -d  --memory 128m nginx
docker container run --name r-memcpu-jenkins -P -d --cpus="0.5" --memory 256m jenkins/jenkins
docker stats

Resource Restrictions in Pods

• Refer Here for official docs
• Limits in Resoruce Restrictions mean maximum size that will be allocated (upper bounds/limits) and request are lower limits
• Exercise: Write the manifest to run above mentioned docker containers in Pod