Spectrumgtw probe is no longer synchronizing inventory only
Article ID: 418052
Updated On:
Products
Issue/Introduction
There are UIM hosts in the Spectrum OneClick console that no longer exist on the UIM side. In the UIM manager hierarchy container, there are duplicate Servers folders. The alarms are synchronized from UIM to Spectrum.
It appears the problem started after updating Spectrum from version 23.3.5 to 24.3.12.
Environment
DX NetOps Spectrum 24.3.12 and DX UIM (Unified Infrastructure Manager) 23.4.5
Cause
UIM required probes for spectrumgtw (Spectrum Gateway) probe were not updated.
In the <UIM_HOME>/probes/gateway/spectrumgtw/vmware.log file:
2025-10-28 18:07:36,226 ERROR SpectrumGtwScheduler_Worker-8: [VMServerJob] [execute] - Catching
com.nimsoft.probe.gateway.spectrumgtw.common.SpectrumGtwException: org.apache.hc.client5.http.HttpResponseException: status code: 500
Resolution
a) Deployed the missing webservices_rest package on the Operator Console (aka UMP) host.
b) Removed the Spectrum UIM Services 8.6.5 package from the primary Hub.
c) Increased the heartbeat_interval from 10000 to 30000 in spectrumgtw probe due to this message error: NAS/EMS or Spectrum is down stopping the schedulers .. NAS/EMS down : true Spectrum down : false
https://knowledge.broadcom.com/external/article?articleId=4872
d) Increased the RAM from 10GB to 26GB on the UMP host.
e) Updated the nisapi_wasp package from 23.4.5 to 23.4.5.1 on the Operator Console (aka UMP) host.
f) Updated the Robot and the other probes on the secondary Hub (where the vmware probe was deployed), as they were on version 23.4.0, however the ppm and prediction_engine probes had already been updated to version 23.4.5. This was the reason why the ppm and prediction_engine probes were not starting and was causing the communication with the spectrumgtw probe.
Feedback
