Downward API implementation for resources limits and requests #24179

aveshagarwal · 2016-04-13T04:10:39Z

This is an implementation of Downward API for resources limits and requests, and it works with environment variables and volume plugin.

This is based on proposal #24051. This implementation follows API with magic keys approach as discussed in the proposal.

@kubernetes/rh-cluster-infra

This change is

aveshagarwal · 2016-05-23T06:27:07Z

Updated PR. Will add more tests soon.
@bgrant0607 @derekwaynecarr @pmorie PTAL.

aveshagarwal · 2016-05-23T18:21:34Z

Added some validation tests and fixed some issues.

aveshagarwal · 2016-05-23T18:26:11Z

I have tested it with following in case any one is interested:

Env variables:

apiVersion: v1
kind: Pod
metadata:
  name: dapi-test-pod
spec:
  containers:
    - name: test-container
      image: gcr.io/google_containers/busybox
      command: [ "/bin/sh", "-c", "env" ]
      resources:
        requests:
          memory: "64Mi"
          cpu: "250m"
        limits:
          memory: "128Mi"
          cpu: "1500m"
      env:
        - name: CPU_LIMIT
          valueFrom:
            resourceFieldRef:
              resource: limits.cpu
    - name: test-container2
      image: gcr.io/google_containers/busybox
      command: [ "/bin/sh", "-c", "env" ]
      resources:
        requests:
          memory: "32Mi"
          cpu: "125m"
        limits:
          memory: "64Mi"
          cpu: "250m"
      env:
        - name: MEMORY_LIMIT
          valueFrom:
            resourceFieldRef:
              resource: limits.memory
        - name: CPU_LIMIT
          valueFrom:
            resourceFieldRef:
              resource: requests.cpu
  restartPolicy: Never

Volume plugin:

apiVersion: v1
kind: Pod
metadata:
  name: kubernetes-downwardapi-volume-example
  labels:
    zone: us-est-coast
    cluster: test-cluster1
    rack: rack-22
  annotations:
    build: two
    builder: john-doe
spec:
  containers:
    - name: client-container
      image: gcr.io/google_containers/busybox
      command: ["sh", "-c", "while true; do if [[ -e /etc/labels ]]; then cat /etc/labels; fi; if [[ -e /etc/annotations ]]; then cat /etc/annotations; fi; sleep 5; done"]
      resources:
        requests:
          memory: "32Mi"
          cpu: "125m"
        limits:
          memory: "64Mi"
          cpu: "250m"
      volumeMounts:
        - name: podinfo
          mountPath: /etc
          readOnly: false
  volumes:
    - name: podinfo
      downwardAPI:
        items:
          - path: "labels"
            fieldRef:
              fieldPath: metadata.labels
          - path: "annotations"
            fieldRef:
              fieldPath: metadata.annotations
          - path: "client-container/pod_name"
            fieldRef:
              fieldPath: metadata.name
          - path: "cpu_limit"
            resourceFieldRef:
              containerName: client-container
              resource: limits.cpu
              divisor: "1"
          - path: "cpu_request"
            resourceFieldRef:
              containerName: client-container
              resource: requests.cpu
              divisor: "1m"
          - path: "memory_limit"
            resourceFieldRef:
              containerName: client-container
              resource: limits.memory
              divisor: "1M"
          - path: "memory_request"
            resourceFieldRef:
              containerName: client-container
              resource: requests.memory

aveshagarwal · 2016-05-23T18:34:44Z

@bgrant0607 @derekwaynecarr @pmorie I have tested it with the above configs and also some other variations including error conditions.

pmorie · 2016-05-23T18:50:57Z

pkg/api/types.go

+	ContainerName string `json:"containerName,omitempty"`
+	// Required: resource to select
+	Resource string `json:"resource"`
+	// Specifies the output format of the exposed resources


Add info about the default here.

pmorie · 2016-05-23T19:27:45Z

@aveshagarwal this is moving in the right direction. Please add e2es for envs and volumes.

aveshagarwal · 2016-05-23T22:10:31Z

@bgrant0607 @derekwaynecarr @pmorie I have updated this PR.

Added e2e tests for env vars, and tested locally and they work.
Improved validation for memory divisor.
Addressed pmorie's comments.

Still working on e2e for volume.

aveshagarwal · 2016-05-23T23:20:06Z

Added e2e tests for volumes too now and tested locally to make sure they work.

@bgrant0607 @pmorie @derekwaynecarr PTAL.

pmorie · 2016-05-24T00:49:28Z

pkg/api/validation/validation.go

+		if downwardAPIVolumeFile.FieldRef != nil {
+			allErrs = append(allErrs, validateObjectFieldSelector(downwardAPIVolumeFile.FieldRef, &validDownwardAPIFieldPathExpressions, fldPath.Child("fieldRef"))...)
+			if downwardAPIVolumeFile.ResourceFieldRef != nil {
+				allErrs = append(allErrs, field.Invalid(fldPath.Child("path"), "resource", "fieldRef and resourceFieldRef can not be specified simultaneously"))


s/.Child("path")//

@aveshagarwal

pmorie · 2016-05-24T00:53:31Z

@aveshagarwal I'll take the nits in a follow-up. LGTM. Thanks a lot.

aveshagarwal · 2016-05-24T11:38:33Z

I see a failure in unit tests: FAIL k8s.io/kubernetes/pkg/api , it seems to be failing in pkg/api/serialization_test.go.

aveshagarwal · 2016-05-24T12:01:40Z

I regenerated auto-generated files and when I test ./hack/test-go.sh pkg/api now it fails in pkg/api/copy_test.go. Not sure what is happening.

aveshagarwal · 2016-05-24T14:49:30Z

I have found the solution and will just run the tests locally and update it and hopefully it should be good to go.

pmorie · 2016-05-24T17:06:38Z

I will re-tag once green

aveshagarwal · 2016-05-24T17:15:36Z

This e2e failure does not seem related to this PR:

[Fail] [k8s.io] Pods [It] should invoke init containers on a RestartNever pod 
/go/src/k8s.io/kubernetes/_output/dockerized/go/src/k8s.io/kubernetes/test/e2e/pods.go:710

aveshagarwal · 2016-05-24T17:19:56Z

Seems similar to this: #26131

pmorie · 2016-05-24T17:20:24Z

@k8s-bot test this issue #26131

aveshagarwal · 2016-05-24T18:20:26Z

@pmorie all green, LGTM again?

aveshagarwal · 2016-05-24T18:22:39Z

@bgrant0607 @derekwaynecarr can you tag it LGTM again?

k8s-github-robot · 2016-05-25T09:46:04Z

@k8s-bot test this [submit-queue is verifying that this PR is safe to merge]

k8s-bot · 2016-05-25T10:25:51Z

GCE e2e build/test passed for commit 1931931.

k8s-github-robot · 2016-05-25T10:38:48Z

Automatic merge from submit-queue

googlebot added the cla: yes label Apr 13, 2016

k8s-github-robot assigned brendandburns Apr 13, 2016

k8s-github-robot added kind/api-change Categorizes issue or PR as related to adding, removing, or otherwise changing an API size/L Denotes a PR that changes 100-499 lines, ignoring generated files. release-note-label-needed labels Apr 13, 2016

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 587f290 to 6fafbea Compare April 14, 2016 02:07

k8s-github-robot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Apr 14, 2016

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 6fafbea to 960e33b Compare April 14, 2016 15:40

k8s-github-robot added needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. and removed needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. labels Apr 14, 2016

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 960e33b to 16dc24c Compare April 18, 2016 17:55

k8s-github-robot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Apr 18, 2016

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 16dc24c to 85bfb54 Compare April 20, 2016 02:40

k8s-github-robot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Apr 21, 2016

aveshagarwal mentioned this pull request Apr 22, 2016

Downward API proposal for resources (cpu, memory) limits and requests #24051

Merged

k8s-github-robot assigned lavalamp and unassigned brendandburns May 6, 2016

lavalamp assigned thockin and unassigned lavalamp May 9, 2016

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 85bfb54 to 54893f1 Compare May 23, 2016 06:25

k8s-github-robot added size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files. and removed size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels May 23, 2016

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 54893f1 to 260841c Compare May 23, 2016 18:20

k8s-github-robot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label May 23, 2016

pmorie added the release-note Denotes a PR that will be considered when it comes time to generate release notes. label May 23, 2016

pmorie reviewed May 23, 2016
View reviewed changes

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 260841c to 55caeb0 Compare May 23, 2016 22:08

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from 55caeb0 to d31acbd Compare May 23, 2016 23:19

aveshagarwal changed the title ~~[WIP] Downward API implementation for resources limits and requests~~ Downward API implementation for resources limits and requests May 23, 2016

pmorie reviewed May 24, 2016
View reviewed changes

pmorie added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label May 24, 2016

Downward API implementation for resources limits and requests

1931931

aveshagarwal force-pushed the master-downward-api-resources-limits-requests-implementation branch from d31acbd to 1931931 Compare May 24, 2016 16:24

k8s-github-robot removed the lgtm "Looks good to me", indicates that a PR is ready to be merged. label May 24, 2016

pmorie added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label May 24, 2016

ncdc added this to the v1.3 milestone May 24, 2016

k8s-github-robot merged commit 83c78fd into kubernetes:master May 25, 2016

aveshagarwal deleted the master-downward-api-resources-limits-requests-implementation branch May 25, 2016 16:14

ncdc mentioned this pull request Jun 3, 2016

Downward API for resource limits #9473

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Downward API implementation for resources limits and requests #24179

Downward API implementation for resources limits and requests #24179

aveshagarwal commented Apr 13, 2016 •

edited

Loading

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

pmorie May 23, 2016

pmorie commented May 23, 2016

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

pmorie May 24, 2016

pmorie commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

pmorie commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

pmorie commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

k8s-github-robot commented May 25, 2016

k8s-bot commented May 25, 2016

k8s-github-robot commented May 25, 2016

Downward API implementation for resources limits and requests #24179

Downward API implementation for resources limits and requests #24179

Conversation

aveshagarwal commented Apr 13, 2016 • edited Loading

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

pmorie May 23, 2016

Choose a reason for hiding this comment

pmorie commented May 23, 2016

aveshagarwal commented May 23, 2016

aveshagarwal commented May 23, 2016

pmorie May 24, 2016

Choose a reason for hiding this comment

pmorie commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

pmorie commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

pmorie commented May 24, 2016

aveshagarwal commented May 24, 2016

aveshagarwal commented May 24, 2016

k8s-github-robot commented May 25, 2016

k8s-bot commented May 25, 2016

k8s-github-robot commented May 25, 2016

aveshagarwal commented Apr 13, 2016 •

edited

Loading