Does multiple model categories lead to ci being reclassified multiple times?

[hyperjam]

My question is next:

Example. I have  LENOVO MODEL X in cmdb_hardware_product_model

Model categories are: Computer, Hardware, OS/X Server etc.

1)Will this lead to situation that discovery once classifies the new CI as Computer, other time while discovery reclassifies as Hardware and etc.? I see this happening in some cases and i wonder if that is the cause, the CI is being bounced between Computer, OS/X Server  and Hardware class. 

I also see for some models that discovery updates the model records with model category Hardware, despite already having Computer category

2) How does discovery identifies which class CI should belong, if there are multiple model categories for 1 model, like: Computer, Hardware, Windows Server, Linux Server

[Fabian Kunzke]

Hello,

Yes and no. A model category is not responsible for reclassification. It "just" tells the system what the model IS used for. For example:

A model has the model category "Computer" and "Server". This means, that a CI, which is a Computer, can have this model assigned. Another CI, which is a Server, can also have this model assigned.

However, it may appear as if the model category dictates the classification. That is not the case. Whenever a model is forcefully assigned (e.g. when discovered) to a class, the matching model category is added. For example:

Let's say you have a model which is only assigned to Computers and only has the model category "Computer". The discovery now finds a "Server" with the same model and assigns it to the CI. Because the CI with that model is of the class Server, the Discovery is like "Oi mate, that model can also be for servers, cause i found one with it!". Thus, the model category "Server" is now also added.

The classification from the discovery is dealt with via the classification engine. E.g. windows classifiers, unix classifiers and so on. The model category has nothing to do with it (or very little). In your case i would highly reccomend to check the ongoing classifications and where the model is "missused". In those cases the classification process for the discovery seems to not be correctly setup leading to the missuse of models.

Hope this helps.
Regards
Fabian