Traps continue forwarding to old SpectroSERVER after Modeling Gateway migration - DX NetOps Spectrum

Products

• Network Observability
• Spectrum

Issue/Introduction

After using Modeling Gateway to migrate or copy device models to a new SpectroSERVER in a distributed environment, traps for those devices may still appear on the original SpectroSERVER. This occurs despite the new models being present and active on the new landscape.

Environment

DX NetOps Spectrum 10.x, 21.x, 22.x
Distributed SpectroSERVER (DSS) environment
Modeling Gateway (MGW)

Cause

The Spectrum Trap Director uses a Remote Forwarding Cache on the Main Location Server (MLS) to route traps to the correct landscape based on the device IP address. When a model is copied via Modeling Gateway, the Trap Director's cache still points to the "Model Handle" on the original SpectroSERVER. The Trap Director continues to route traps to the original location until that cache entry is invalidated.

Resolution

The Remote Forwarding Cache is updated when the original model is deleted or when the cache entry expires.

1. Delete Original Models: Remove the duplicate device models from the original SpectroSERVER landscape. This action notifies the Trap Director to clear the old model handle from its cache.
2. Verify New Model State: Ensure the models on the new SpectroSERVER have transitioned from "Blue" (Initial Condition) to "Green" (Normal). The SpectroSERVER must successfully poll the device via SNMP or ICMP to establish management.
3. Manual Cache Refresh: If traps do not immediately cut over after deleting the old models, manually clear the cache on the MLS:
   • Locate the VNM model on the Main Location Server (MLS).
   • In the Component Detail panel, go to Information > Trap Management > Remote Forwarding Cache.
   • Select the entries for the impacted IPs and remove them, or click Clear Cache to force a global "Distributed Find" for all incoming traps.