Infinite deployment repair loop caused by DB–k8s state mismatch due to persistent storage name collisions #690
Description
Deployments enter an infinite repair cycle because the deployment state stored in the database differs from the state in k8s.
Root Cause
Persistent storage entries that share the same name overwrite each other. This results in an incorrect persisted deployment state, which continuously differs from the actual k8s deployment and triggers repeated repairs.
Proposed Solution
Enforce unique names for persistent storage entries to prevent overwriting and ensure consistency between the database and k8s state.