Want to use disasterRecovery: for etcd

[navneetrajput06]

Hi team,
We are trying to use for etcd using below section from etcd config. but for cronjob its picking wrong API version.
disasterRecovery:
enabled: true

Could you please help how can it be passed correct version.
error: unable to recognize "STDIN": no matches for kind "CronJob" in version "batch/v1beta1"

[yhmo]

I suppose it is a k8s-related question, you can search for more info on google or k8s community. Maybe the version should be "apiVersion: batch/v1"

[navneetrajput06]

@yhmo yes its k8s-related question but we dont have control on API version for Kind Cronjob. its trying to deploy with etcd configuration. when we try to deploy any other cronjob we are using batch/v1 and thats working. but this is being deployed using etcd dynamically via Milvus.yaml. There is no option to pass API version.

[LoveEachDay]

@navneetrajput06 How do you create an etcd cluster? The Milvus Helm chart currently uses an outdated version of the etcd chart, which relies on an old CronJob API that is hardcoded in the embedded chart. This means the API version cannot be changed via values.yaml.

If you need to enable disaster recovery for etcd, we recommend deploying an external etcd cluster using the Bitnami etcd chart instead.

[navneetrajput06]

@LoveEachDay Got your point. Do we have any plan to add chart versions configurable ?

[navneetrajput06]

@LoveEachDay @yhmo we were able to deploy using flux and able to change API version in template and cron job was deployed. but snapshotting pod is going in crashloopOff and throwing below error in logs.
etcd ERROR ==> all etcd endpoints are unhealthy!

Below is the etcd yaml added into milvus.
disasterRecovery:
enabled: true
cronjob:
schedule: "*/30 * * * *"
pvc:
existingClaim: "etcd-snapshot-pvc"