[FLINK-33222] Operator rollback app when it should not

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Kubernetes Operator

Environment

Flink operator 1.6 - Flink 1.17.1

• Assignee: Unassigned
• Reporter: Nicolas Fraison
• Created: 09/Oct/23 13:35
• Updated: 10/Oct/23 08:54

Description

The operator can decide to rollback when an update of the job spec is performed on 
savepointTriggerNonce or initialSavepointPath if the app has been deployed since more than KubernetesOperatorConfigOptions.DEPLOYMENT_READINESS_TIMEOUT.
 
This is due to the objectmeta generation being updated when changing those spec and leading to the lastReconcileSpec not being aligned with the stableReconcileSpec while those spec are well ignored when checking for upgrade diff
 
Looking at the main branch we should still face the same issue as the same update is performed at the end of the reconcile loop