Kubernetes show Cron Job Successfully but when not getting desired result

Question

when I run the cronjob into Kubernetes, that time cron gives me to success the cron but not getting the desired result

apiVersion: batch/v1beta1
kind: CronJob
metadata:
  name: {{ $.Values.appName }}
  namespace: {{ $.Values.appName }}
spec:
  schedule: "* * * * *"
  concurrencyPolicy: Forbid
  jobTemplate:
    spec:
      template:
        spec:
          containers:
            - name: test
              image: image
              command: ["/bin/bash"] 
              args: [ "test.sh" ]
          restartPolicy: OnFailure

also, I am sharing test.sh

#!/bin/sh
rm -rf /tmp/*.*
echo "remove done"

cronjob running successfully but when I checked the container the file is not getting deleted into /tmp directory

Doesn't this just delete a directory that does not exists in the container? Do you want to mount a volume where this directory exists? — Jonas, May 27 '21 at 17:01
Yes I added volume into /tmp directory, I need to delete all file and directory from /tmp — Parth Shah, May 28 '21 at 04:44
Hi, please update your `YAML` manifest with the `Volume` mounting that you've referenced under one of the answers. — Dawid Kruk, May 28 '21 at 14:04

score 2 · Answer 1 · answered May 28 '21 at 05:32

Cronjob run in one specific container, if you want to remove the file or directory from another container it won't work.

If your main container running under deployment while when your job or cronjob gets triggered it create a new container (POD) which has a separate file system and mount option.

if you want to achieve this scenario you have to use the PVC with ReadWriteMany where multiple containers (POD) can connect with your single PVC and share the file system.

In this way then your cronjob container (POD) get started with the existing PVC file system and you can remove the directory using job or cronjobs.

mount the same PVC to the cronjob container and the main container and it will work.

yes I have attached and mount efs volume on /tmp the but the issue is echo command is getting executed but fails to execute "rm" — Parth Shah, May 28 '21 at 06:14
you have mount efs volume on /tmp but I can not see in your cronjob YAML it's getting attach to /tmp ? efs as PVC are you attaching? — Harsh Manvar, May 28 '21 at 08:56

score 2 · Accepted Answer · answered Jun 01 '21 at 09:58

You need to have the persistence volume attached with you pod and cronjob that you are using so it can remove all the files when the script get executed. You need to mount and provide path accordingly in your script. For adding kubernetes cronjobs kindly go through this link

score 1 · Answer 3 · answered May 27 '21 at 18:47

Change test.sh to:

#!/bin/sh

set -e

rm -rf /tmp/*.*
echo "remove done"

Without -e your bash script will return with the same status as its last command. In this case, it's an echo so it will always have status 0 (success). Using set -e will make your script abort and fail if the rm command fails.

Also, without any volume mounts, this cron job does not do anything meaningful. If you want to delete some files from another container, you would need to use Cron within that container (or have a volume with ReadWriteMany).

Kubernetes show Cron Job Successfully but when not getting desired result

3 Answers3