Configure the ROBOT_NIS_CACHE Subject queue in DX UIM 23.4

Products

• DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

What is ROBOT_NIS_CACHE?

Why is this subject a requirement in DX UIM 23.4.0 and higher?

Why should I create a ROBOT_NIS_CACHE queue?

Environment

DX UIM 23.4.0 and higher

Cause

Changes to robot discovery process

Resolution

Starting with Discovery Server Probe v23.4.0, discovery does not 'pull' the inventory information like monitoring devices, configuration_items, and metrics lists from robots v23.4.0 or higher. Instead, the discovery server listens to the new hub queue with the subject ROBOT_NIS_CACHE on the Primary hub. The secondary hub PUSHES the robot nis_cache contents to the discovery server for processing. The discovery server processes the content from the newly created queue.

If the queue is not configured new devices may not showing metrics in the OC views although they are correctly storing data in the DB.

You can read more about how this functionality works Configure Discovery Queues (broadcom.com)

 

• Key points are as follows:

  • In prior versions of DX UIM, robots were scanned by discovery_server using a "pull" methodology, where the discovery_server built a list of robots in the environment and requested the contents of the niscache in order to discover the robots.

  • In UIM 23.4.0 and higher, when a robot (controller) is also version 23.4.0 or higher, this scan will not be performed, but instead, the controller now uses a "push" method to publish niscache changes.

  • To enable this functionality you must either add a new queue (this approach is preferred), or modify the message Subject(s) on an existing queue, to push messages with ROBOT_NIS_CACHE Subject up to the Primary hub.

  • Until the queue(s) exist, the robot will not publish changes to niscache, and new robots installed at version 23.4.0 or higher will not be discovered.

  • Older robots (prior to 23.4.0) will still be scanned individually using the traditional "pull" method.

  • It is recommended to use an ATTACH/GET queue combination as opposed to a POST queue, as a POST queue may cause undesired side-effects in an environment with hub tunnels. For example set up a queue on a downstream hub <hubname1>_robot_nis_cache, configured as an ATTACH queue and setup the corresponding GET queue on the Primary hub. Set the GET queue bulk_size as 1000. The existing robot_nis_cache attach queue which is installed on the Primary hub during the upgrade can be left as is.

 

• How to create the ROBOT_NIS_CACHE queue: 

1. 1. Go to the secondary (downstream) hub and open the hub GUI, go to queues, and create new queue. 
      
      
   2. Create new attach queue with a preferred name, and select from the dropdown the subject ROBOT_NIS_CACHE
      
      
   3. Go to the primary hub and create a new GET queue, select the secondary hub from the drop down and select the queue created for ROBOT_NIS_CACHE
      
      
   4. After the hubs have automatically restarted, the queues will connect. 

Additional Information

niscache - what it is, what it does