Scribd Logo
Upload

Welcome to Scribd!

  • Lab 6

    Uploaded by

    Rajasekhar Reddy
    186 views4 pages
    Lab6

    Original Title

    Lab6

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    Lab6

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    Download as pdf or txt
    186 views4 pages

    Lab 6

    Uploaded by

    Rajasekhar Reddy
    Lab6

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    Download as pdf or txt
    of 4

    1

    _______________________________________________________________________________________________________________

    LAB 6

    Objective​: Create a deployment, scale it, trigger a rolling update and rollback. Understand how replica
    sets allow you to manage revisions.

    Pods are the lowest compute unit of Kubernetes. To ensure that pods are always running, you define a
    deployment​. A deployment is a declarative manifest that describes what Pods should be running and
    how many.

    To create your first deployment, use the ​kubectl run ​command. It is a convenience wrapper to
    ​
    define single container Pods. Let’s create a deployment for ​Ghost, the microblogging platform:

    $ kubectl run ghost --image=ghost


    $ kubectl get deployments
    NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
    ghost 1 1 1 1 6s

    With your deployment running, check that a corresponding Pod has been started:

    $ kubectl get pods


    NAME READY STATUS RESTARTS AGE
    ghost-943298627-kdhts 1/1 Running 0 9s

    Now, scale your deployment. For example, let’s scale it to five replicas:

    $ kubectl scale deployments ghost --replicas=5


    $ kubectl get pods
    NAME READY STATUS RESTARTS AGE
    ghost-943298627-38761 0/1 ContainerCreating 0 2s

    ____________________________________________________________________________________________________________

    LFS258 Copyright 2017. The Linux Foundation. All rights reserved.​ ​


    2
    _______________________________________________________________________________________________________________

    ghost-943298627-ghshb 0/1 ContainerCreating 0 2s


    ghost-943298627-kdhts 1/1 Running 0 4m
    ghost-943298627-slzdr 0/1 ContainerCreating 0 2s
    ghost-943298627-xwljz 0/1 ContainerCreating 0 2s

    To set the desired number of Pods, a replica sets resource has been automatically generated by the
    deployment. You can list this replica set with ​kubectl get rs​. It continuously reconciles the desired
    state (e.g five replicas) with the actual cluster state. The replica set counts the number of Pods that
    matches the Pod selector defined in the deployment manifest and scales the Pods accordingly. Below,
    ​ q​ to list the pod selection via labels.
    we show how you can parse the deployment manifest with j

    $ kubectl get deployment ghost -o json | jq -r .spec.selector


    {
    "matchLabels": {
    "run": "ghost"
    }
    }

    To see that the replica set does its job, delete one of the pods, and list Pods again. You will see that
    another one got started to still have the desired number of replicas running. You can also see this by
    removing the label from one of the running pods:

    $ kubectl label pods ghost-943298627-38761 run-


    $ kubectl get pods -Lrun
    NAME READY STATUS RESTARTS AGE RUN
    ghost-943298627-38761 1/1 Running 0 1h <none>
    ghost-943298627-ghshb 1/1 Running 0 1h ghost
    ghost-943298627-kdhts 1/1 Running 0 1h ghost
    ghost-943298627-nj122 1/1 Running 0 56m ghost
    ghost-943298627-slzdr 1/1 Running 0 1h ghost
    ghost-943298627-xwljz 1/1 Running 0 1h ghost

    ____________________________________________________________________________________________________________

    LFS258 Copyright 2017. The Linux Foundation. All rights reserved.​ ​


    3
    _______________________________________________________________________________________________________________

    Above you see that one Pod does not have the label ​run=ghost​, because we removed it. The replica
    set automatically started a new Pod to have five running.

    Explore a bit more the ​kubectl labels command and check how they can be set in resource
    metadata.

    Rolling updates and rollbacks

    One big advantage of deployments is that they can be used to do a rolling update of your Pods. For
    example, let’s assume that you have a new image that you want to use. You can use the ​kubectl
    set​ command to change that image.

    $ kubectl set image deployment/ghost ghost=ghost:0.9

    List the replica sets again and see that a new set was created. This new set corresponds to the new
    image. The old and new replica sets will be scaled down and up respectively to keep the application
    running but provide the new image. At the end of the update, the new replica set has five running
    replicas and the old one has zero.

    $ kubectl get rs
    NAME DESIRED CURRENT READY AGE
    ghost-3487275284 2 2 0 4s
    ghost-943298627 4 4 4 1h
    kubectl get rs
    NAME DESIRED CURRENT READY AGE
    ghost-3487275284 5 5 4 1m
    ghost-943298627 0 0 0 1h

    If you list your pods, you will see that the running pods now use the new replica set, except the pod
    that we re-labeled and was taken out of the replica set:

    $ kubectl get pods


    NAME READY STATUS RESTARTS AGE
    ghost-3487275284-73lfn 1/1 Running 0 1m
    ghost-3487275284-8pxvb 1/1 Running 0 2m

    ____________________________________________________________________________________________________________

    LFS258 Copyright 2017. The Linux Foundation. All rights reserved.​ ​


    4
    _______________________________________________________________________________________________________________

    ghost-3487275284-pcbs5 1/1 Running 0 1m


    ghost-3487275284-pnw4j 1/1 Running 0 2m
    ghost-3487275284-t2b4q 1/1 Running 0 1m
    ghost-943298627-38761 1/1 Running 0 1h

    Now, check the history of your deployment:

    $ kubectl rollout history deployment/ghost


    deployments "ghost"
    REVISION CHANGE-CAUSE
    1 <none>
    2 <none>

    You will see two revisions. The change-cause will be empty. You can record that cause by specifying
    the ​--record​ option when you create your deployment.

    To rollback, simply do:

    $ kubectl rollout undo deployment/ghost

    If you keep an eye on your Pods, you will see that they will be terminated and new ones corresponding
    to the original replica set will be created. To learn more about deployments, do not forget to check the
    documentation​.

    ____________________________________________________________________________________________________________

    LFS258 Copyright 2017. The Linux Foundation. All rights reserved.​ ​

    You might also like