Run in kubernetes as a cron job using a service account

Here’s how to set up a cluster backup using a CronJob and a service account.

You can read about CronJobs in doc

Create namespace

First, let’s create a namespace kube-dump (if you want to use a different name, you will need to edit it in all the necessary manifests)

kubectl create ns kube-dump

Deploy serviceaccount

⚠️ Attention! ⚠️

This is an example to demonstrate how it works.
Do not do this in a production environment, the example below uses the cluster view role cluster-role-view.yaml, you have enough view permissions to dump the data.

Configure your role correctly according to your needs. You can find configuration examples in the /deploy directory.

Apply cluster admin sericeaccount, pvc, pod in kube-dump namespace:

kubectl apply -n kube-dump -f deploy/cluster-role-view.yaml

Deploy with git repository oauth token

If you want to use a git repository to store dumps, you will need to provide the address of the remote repository and credentials.

Let’s create a secret with the GIT_REMOTE_URL variable in which we will specify the https repository address. Insert your oauth token in place of the $TOKEN variable:

kubectl -n kube-dump create secret generic kube-dump \
  --from-literal=GIT_REMOTE_URL=https://oauth2:$TOKEN@corp-gitlab.com/devops/cluster-bkp.git

And apply the cron job manifest:

kubectl apply -n kube-dump -f deploy/cronjob-git-token.yaml

Deploy with git repository write allowed ssh key

Generate ssh key:

mkdir -p ./.ssh
chmod 0700 ./.ssh
ssh-keygen -t ed25519 -C "kube-dump" -f ./.ssh/kube-dump
cat ./.ssh/kube-dump.pub

Show public ssh key and add to repository deployment keys with write access.

cat ./.ssh/kube-dump.pub

Create pvc for store data such as cache

kubectl apply -n kube-dump -f deploy/pvc.yaml

Create secret with private ssh key

kubectl -n kube-dump create secret generic kube-dump-key \
  --from-file=./.ssh/kube-dump \
  --from-file=./.ssh/kube-dump.pub

And apply the cron job manifest, previously you could set up environment variables

kubectl apply -n kube-dump -f deploy/cronjob-git-key.yaml

Deploy with persistent volume to store achives

If you want to use archives saved to disk for storing dumps, you will need to create a PVC and connect it to a cron job.

kubectl apply -n kube-dump -f deploy/pvc.yaml

And apply the cron job manifest:

kubectl apply -n kube-dump -f deploy/cronjob-archives.yaml