bambash/helm-cronjobs

a helm chart for cronjobs in kubernetes

helm-cronjobs

You can define an array of jobs in values.yaml helm will take care of creating all the CronJobs.

How to use as a starter chart

1. Find your Helm data directory, HELM_DATA_HOME

   helm env
   

2. cd to this directory, then

   mkdir starters
   cd starters
   

3. Clone this repo

4. In your cronjob project, set up your new chart with

   helm create -p helm-cronjobs your_chart_name
   

Configuration

Via values.yaml

Overview

jobs:
  jobname-1:
    # job definition
  jobname-2:
    # job definition
  jobname-n:
    # job definition

Details

jobs:
  ### REQUIRED ###
  <job_name>:
    image:
      repository: <image_repo>
      tag: <image_tag>
      imagePullPolicy: <pull_policy>
    schedule: "<cron_schedule>"
    failedJobsHistoryLimit: <failed_history_limit>
    successfulJobsHistoryLimit: <successful_history_limit>
    concurrencyPolicy: <concurrency_policy>
    restartPolicy: <restart_policy>
  ### OPTIONAL ###
    imagePullSecrets:
    - username: <user>
      password: <password>
      email: <email>
      registry: <registry>
    env:
    - name: ENV_VAR
      value: ENV_VALUE
    envFrom:
    - secretRef:
      name: <secret_name>
    - configMapRef:
      name: <configmap_name>
    command: ["<command>"]
    args:
    - "<arg_1>"
    - "<arg_2>"
    resources:
      limits:
        cpu: <cpu_count>
        memory: <memory_count>
      requests:
        cpu: <cpu_count>
        memory: <memory_count>
    serviceAccount:
      name: <account_name>
      annotations:  # Optional
        my-annotation-1: <value>
        my-annotation-2: <value>
    nodeSelector:
      key: <value>
    tolerations:
    - effect: NoSchedule
      operator: Exists
    volumes:
      - name: config-mount
        configMap:
          name: configmap-name
          items:
            - key: configuration.yml
              path: configuration.yml
    volumeMounts:
      - name: config-mount
        mountPath: /etc/config
    affinity:
      nodeAffinity:
        requiredDuringSchedulingIgnoredDuringExecution:
          nodeSelectorTerms:
          - matchExpressions:
            - key: kubernetes.io/e2e-az-name
              operator: In
              values:
              - e2e-az1
              - e2e-az2

Examples

$ helm install test-cron-job .
NAME:   cold-fly
LAST DEPLOYED: Fri Feb  1 15:29:21 2019
NAMESPACE: default
STATUS: DEPLOYED

RESOURCES:
==> v1/CronJob
NAME                    AGE
cold-fly-hello-world    1s
cold-fly-hello-ubuntu   1s
cold-fly-hello-env-var  1s

list cronjobs:

$ kubectl get cronjob
NAME                     SCHEDULE      SUSPEND   ACTIVE    LAST SCHEDULE   AGE
cold-fly-hello-env-var   * * * * *     False     0         23s             1m
cold-fly-hello-ubuntu    */5 * * * *   False     0         23s             1m
cold-fly-hello-world     * * * * *     False     0         23s             1m

list jobs:

$ kubectl get jobs
NAME                                DESIRED   SUCCESSFUL   AGE
cold-fly-hello-env-var-1549056600   1         1            45s
cold-fly-hello-ubuntu-1549056600    1         1            45s
cold-fly-hello-world-1549056600     1         1            45s

helm-cronjobs's Issues

Deploy configmaps with the chart

Several cronjobs require a configuration file.

With the current chart, we need to deploy the configmap manually and use it in the chart.

It will be nice if the chart is able to create the configmap from a list of files.

I have no experience with chart development, but if you agree on the feature I can try to implement it

Could you add a license?

Title says it all

nil pointer evaluating interface {}.repository

Hi,
Creating a brand new cron job works fine, but when helm upgrade --install is called to update an existing cron job I see the following error....

helm.go:76: [debug] template: app/charts/datacollector/templates/cronjob.yaml:44:27: executing "app/charts/datacollector/templates/cronjob.yaml" at <$job.image.repository>: nil pointer evaluating interface {}.repository

my values.yaml looks like this...

  jobs:
  - name: pdschedule
    image:
      repository: private/pdschedule
      tag: 2
      imagePullPolicy: IfNotPresent
    schedule: 0 */4 * * *
    failedJobsHistoryLimit: 1
    successfulJobsHistoryLimit: 3
    concurrencyPolicy: Forbid
    restartPolicy: OnFailure
    command:
    - /bin/sh
    args:
    - -c
    - java $JAVA_OPTS -jar /app.jar com.example.app

An error occurs when deploying this chart

Run command on gitlab runner:

$ helm upgrade --install "${COMPONENT}" chartmuseum/helm-cronjobs --version "${API_BACKEND_CHART_VERSION}" --namespace "${CI_K8S_NAMESPACE}" --values ./deploy/${VALUES}.yaml --wait --timeout "${HELM_TIMEOUT}"

When deploying to a dev environment in k8s, an error occurs:

coalesce.go:175: warning: skipped value for helm-cronjobs.jobs: Not a table.
Error: UPGRADE FAILED: template: helm-cronjobs/templates/serviceaccount.yaml:6:21: executing "helm-cronjobs/templates/serviceaccount.yaml" at <$job>: wrong type for value; expected map[string]interface {}; got string

How can this error be fixed?

Deploy service accounts

Several cronjobs require a service account to run correctly.

With the current chart, they are not supported.

It will be nice if the chart is able to create the service account and attach it to the jobs.

I have no experience with chart development, but if you agree on the feature I can try to implement it