SystemEDGE manager_name points to non-resolving hostname, causing recurring alarm

Products

• DX OI SaaS

Issue/Introduction

We're receiving repeated "Management Agent Lost" alarms in Spectrum for one of our servers

The alarms keep triggering and clearing, and we're not sure what's causing them. The server itself appears to be up and reachable (we can ping it), and system resources (CPU/memory/disk) look normal.

Environment

Environment: Broadcom DX NetOps Spectrum. SystemEDGE Version 5.9.0 Build 15037.

Cause

Server was rebooted to correct a Sysedge issue, however the Spectrum Data Publisher did not come up correctly.  At start time it could not find some libraries. The process was restarted as root which compounded the issue.

Resolution

Restarted the process as user spectrum and all reporting correctly