Skip to content

Latest commit

 

History

History
60 lines (44 loc) · 1.86 KB

resources.md

File metadata and controls

60 lines (44 loc) · 1.86 KB
Raw

Resources

The best way to find the documentation is narrowing down the resources by using kk explain

kk explain pod.spec.containers | grep -i resources
---
resources	<Object>
Compute Resources required by this container. Cannot be updated. More info:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/

As you can see, the resources are defined within the specific pod's container.

There are two components

  1. Requests
  2. Limits

In layman's terms (aka me), the kube-scheduler is going to take these requests and limits and search for a node that has the available requested capacity (whether it be disk space or compute power).

The documentation does an excellent job running this down in detail. I'd highly suggest reading it in more detail.

Adding requests and limits

Let's make a pod and add some resource requests and limits

apiVersion: v1
kind: Pod
metadata:
  name: p1
spec:
  containers:
  - image: busybox
    args:
    - /bin/sh
    - -c
    - sleep 3600
    name: p1
    resources:
      requests:
        memory: "64Mi"
        cpu: "100M"
      limits:
        memory: "128Mi"
        cpu: "200M"

If the pod fails to be assigned to a node, it is relatively easy to debug. The pod will sit in the pending state forever until additional capacity is added.

NAME   READY   STATUS    RESTARTS   AGE
p1     0/1     Pending   0          35s

Take a look at the pod description too to see the error

kk describe po p1 | grep -i events -A 10
---
Events:
  Type     Reason            Age                From               Message
  ----     ------            ----               ----               -------
  Warning  FailedScheduling  4s (x20 over 94s)  default-scheduler  0/2 nodes are available: 2 Insufficient cpu.