Option for persistentVolumeClaimRetentionPolicy

[greenlama]

Is this a BUG REPORT or FEATURE REQUEST? (choose one):
FEATURE REQUEST

Version of Helm and Kubernetes:
kubernetes v1.26

Which chart:
stable/artifactory-ha

Which product license (Enterprise/Pro/oss):
Enterprise

What happened:
artifactory-primary-statefulset doesn't have persistentVolumeClaimRetentionPolicy

What you expected to happen:
artifactory-primary-statefulset has persistentVolumeClaimRetentionPolicy and can be managed by editing at Values.yaml. persistentVolumeClaimRetentionPolicy is needed since PVC is being created within the availability zone. This affects when pods can be restarted in different availability zone. So we need to have an option to delete PVC on pod destroy.

[RobinDuhan]

Hi @greenlama, thanks for clarifying. My understanding is that the pod could get stuck waiting if it schedules in a different zone than its PVC. This seems like an infrastructure problem, potentially solvable with a single-zone node pool combined with tolerations for Artifactory. Node affinity is another option, guiding Artifactory to nodes labeled zone=ap-south-1a (for instance); with a large enough pool, nodes usually exist in each zone.

I understand your point about persistentVolumeClaimRetentionPolicy. My concern there is if the PVC gets deleted and recreated upon restart, we might lose the benefit of statefulness, which could cause issues.

[RobinDuhan]

Since this is not a chart issue. I am closing this. Please re-open if needed.