When I want to create an SPM test, I am easily able to discover a systemEdge agent running on another SS, but to discover an agent installed on itself (to use as an SPM testHost), is more of a challenge.
If an agent is installed on the same SS, that is then used to discover and model it, duplicate IP errors will appear on the model, and an instability may be experienced on the SPM testHost. e.g. The VNM model, with the same IP address, can replace the SystemEDGE agent as testHost and limits available test to ICMP PING tests only. The ability to use can be affected also e.g. the ADES (caesadaim) plugin is not supported with the VC AIM.
There are however workarounds to avoid such problems, which will not all will work on every system e.g. depending on the type of network configuration used.
1. If the SpectroSERVER has a valid IPv6 address and is on an IPv6 network, we can then use the IPv4 address to discover the agent.
2. We can discover the agent using the loopback address. Using 127.0.0.1 to discover a new model is not always possible, so creating a new "model by hostname" using “localhost” which resolves to 127.0.0.1 will avoid this problem.
Both of these options avoid duplication errors on the model, the address will remain unchanged after an SS restart and the testHost should remain stable.
Please note that changing the network_address of the VNM model to 127.0.0.1 and discovering the agent using the standard address will function correctly until we restart the SS, at which point the VNM address will revert back to its standard IP address, giving us duplicate IP address errors on the model and once again instability on the testHost.
It is recommended that we only install Spectrum software on a SpectroSERVER and not to use the SpectroSERVER for purposes other than originally intended so these recommendations are for those who have no other choice in doing so. This has not been tested by QA and is therefore not officially supported by CA Spectrum.