Resource Quota Allocation

Resource Quota Allocation

Create a new namespace:

kubectl create ns limited-cpu-memory

namespace/limited created

Now create a new file:

vi /home/student/compute-quota.yaml

File: /home/student/compute-quota.yaml

apiVersion: v1
kind: ResourceQuota
metadata:
  name: compute-quota
  namespace: limited-cpu-memory
spec:
  hard:
    requests.cpu: "1"
    requests.memory: 1024Mi
    limits.cpu: "5"
    limits.memory: 2048Mi

Create the compute-quota:

kubectl create -f compute-quota.yaml

resourcequota/compute-quota created

Check the details of the compute-quota:

kubectl describe quota -n limited-cpu-memory

Name:            compute-quota
Namespace:       limited
Resource         Used  Hard
--------         ----  ----
limits.cpu       0     5
limits.memory    0     2Gi
requests.cpu     0     1
requests.memory  0     1Gi

Similarly, we can define quotas for other Kubernetes objects with the following ResourceQuota object.

Create a new namespace:

kubectl create ns limited-count-object

namespace/limited-count-object created

Create a new file:

vi object-quota.yaml

File: /home/student/object-quota.yaml

apiVersion: v1
kind: ResourceQuota
metadata:
  name: object-counts
  namespace: limited-count-object
spec:
  hard:
    configmaps: "5"
    persistentvolumeclaims: "4"
    pods: "10"
    secrets: "10"
    services: "10"
    services.nodeports: "2"

Apply these object quotas:

kubectl create -f object-quota.yaml

resourcequota/object-quota created

Check on the object-counts object status:

kubectl describe quota object-counts -n limited-count-object

Name:                   object-counts
Namespace:              limited-count-object
Resource                Used  Hard
--------                ----  ----
configmaps              0     5
persistentvolumeclaims  0     4
pods                    0     10
secrets                 1     10
services                0     10
services.nodeports      0     2

At this point, try to create a simple deployment:

kubectl create deployment test01 --image=r.deso.tech/library/nginx -n limited-count-object

deployment.apps/test01 created

Scale your test01 deployment to 11 replicas:

kubectl scale deployment test01 --replicas=11 -n limited-count-object

deployment.apps/test01 scaled

Now, check your environment and investigate what happened:

kubectl get deploy,rs,pod -n limited-count-object

NAME                     READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/test01   10/11   10           10          7m13s

NAME                                DESIRED   CURRENT   READY   AGE
replicaset.apps/test01-67f6859667   11        10        10      40s

NAME                          READY   STATUS    RESTARTS   AGE
pod/test01-67f6859667-44rdg   1/1     Running   0          40s
pod/test01-67f6859667-6jf2j   1/1     Running   0          40s
pod/test01-67f6859667-f52d8   1/1     Running   0          40s
pod/test01-67f6859667-gfktj   1/1     Running   0          40s
pod/test01-67f6859667-h7nrh   1/1     Running   0          40s
pod/test01-67f6859667-krgjj   1/1     Running   0          39s
pod/test01-67f6859667-lhxdk   1/1     Running   0          40s
pod/test01-67f6859667-lxvgz   1/1     Running   0          40s
pod/test01-67f6859667-thpwd   1/1     Running   0          39s
pod/test01-67f6859667-vw5n5   1/1     Running   0          39s

Where are your 11 pods? Investigate your deployment and replicaset:

kubectl describe deploy test01 -n limited-count-object

kubectl describe rs test01-xxxxxx -n limited-count-object

If you look in the events, you’ll see some errors. Kubernetes can’t schedule all your pods because you reached the maximum defined in ResourseQuota.

Check on the object-counts object status:

kubectl describe quota object-counts -n limited-count-object

Name:                   object-counts
Namespace:              limited-count-object
Resource                Used  Hard
--------                ----  ----
configmaps              0     5
persistentvolumeclaims  0     4
pods                    10    10
secrets                 1     10
services                0     10
services.nodeports      0     2

Clean up

kubectl  delete ns limited-cpu-memory limited-count-object

namespace "limited-cpu-memory" deleted
namespace "limited-count-object" deleted