# k8s Operator Restart after machine fail

**URL:** <https://forums.percona.com/t/k8s-operator-restart-after-machine-fail/7352>\
**Category:** Percona XtraDB Cluster 5.x\
**Created:** [December 16, 2019, 3:39am UTC](https://forums.percona.com/t/k8s-operator-restart-after-machine-fail/7352 "2019-12-16T03:39:23Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![ewaldrg](https://avatars.discourse-cdn.com/v4/letter/e/94ad74/32.png) [@ewaldrg](https://forums.percona.com/u/ewaldrg)\
**Post date:** [December 16, 2019, 3:39am UTC](https://forums.percona.com/t/k8s-operator-restart-after-machine-fail/7352/1 "2019-12-16T03:39:23Z")

</div>

Hi,

while testing the k8s operator (with 3 pods each) in my Minikube environment i had to restart Minikube.  
After that the cluster didn’t come back up. (CrashLoopBackOff)

In our current, non-k8s-environment i made a script to force bootstrap in such a situation, keeping in mind that data could be lost:  
Setting the “safe\_to\_bootstrap” to 1 in grastate.dat and call “service mysql bootstrap-pxc” on the selected node.

So i ran a Job that sets this variable in the PVC of the first pod:

```auto

apiVersion: batch/v1
kind: Job
metadata:
name: safe-to-bootstrap
spec:
template:
spec:
volumes:
- name: mysql-data
persistentVolumeClaim:
claimName: datadir-cluster1-pxc-0
containers:
- name: safe-to-bootstrap
image: busybox
imagePullPolicy: IfNotPresent
command:
- sed
- -i
- "s|safe_to_bootstrap.*:.*|safe_to_bootstrap:1|1"
- /var/lib/mysql/grastate.dat
volumeMounts:
- mountPath: /var/lib/mysql
name: mysql-data
restartPolicy: OnFailure

```

Then i deleted all the “cluster1-pxc-” pods to bring it back up.

Is there a safer and/or cleaner way to do this?