Discovery over Load Balancer or clustered IP addresses

  • KM02538294
  • 30-Aug-2016
  • 07-Nov-2016

Summary

Best practice on how to run UCMDB Discovery over clustered IPs

Question

Application_IP attribute switches between node IP and clustered IP assigned to the CRG (ClusterResourceGroup) on which the node resides.

Answer

When a Host Connection by Shell or Host Application by Shell runs over a clustered IP address then all the RunningSoftware can have the attribute application_ip switching between the clustered IP and the IP of the node part of the cluster.

This is due to the discovery logic which can't tell what the primary IP is (the node will have multiple IP's assigned as it's part of a cluster or defined in a Load Balancer) of the node and it will report the application IP as the IP used to connect to that particular trigger node.

UCMDB R&D's suggestion is to exclude from the Probe IP ranges all the clustered IPs and IPs defined on Load Balancer VIPs. This will avoid dispatching discovery jobs on the clustered IPs and thus the reported data will always be accurate.