Loading

RTF Apps Still Running Old Images / Control Plane Out of Sync After an RTF Upgrade

Date de publication: Jun 24, 2026
Description

PROBLEM

After upgrading Runtime Fabric (RTF), one or more deployed applications continue running on the previous image version. The Anypoint Runtime Manager UI may also appear out of sync, showing stale or inconsistent application state.

ROOT CAUSE

Kubernetes does not automatically replace running pods when a new image becomes available on the node. Pods that were healthy at the time of the RTF upgrade are left in place. They continue running on the old image until they are explicitly recycled. This is expected Kubernetes behavior, not a defect in RTF.

Control plane synchronization delays are a separate concern and resolve automatically once the upgrade completes; no manual intervention is required for the UI to catch up.

Résolution

SOLUTION

1. Recycle Affected Applications

For each application still running the old image, perform a Stop followed by a Start from Anypoint Runtime Manager (or via the CLI). This forces Kubernetes to schedule a new pod, which pulls the current image from the registry.

- Stop the application and wait for it to reach a stopped state before starting it again.

- Repeat for each affected application.

2. Control Plane Sync

No action is required. The Runtime Manager UI will re-sync with cluster state automatically after the upgrade. If the UI remains inconsistent after recycling applications, allow a few minutes for the control plane to reconcile before investigating further.

Confirm that the restarted applications report the expected image version and that Runtime Manager reflects the correct running state.

Numéro d’article de la base de connaissances

005387639

 
Chargement
Salesforce Help | Article