Discovery Process Classifiers creating duplicates

[john_duchock]

While i think i have seen an article on my problem, the past two days of searching have gone in vain...so maybe someone knows and i am looking in the wrong place (points to Doug)...

I have a simple process classifier that creates a record on cmdb_ci_appl based on an executable running (provenirADMIN.exe).   Setting this up was easy enough, but when i discover two servers that have the process running, it creates unique values.   These are typically:

provenirADMIN.exe@servername1   Runs on::Runs   servername1

provenirADMIN.exe@servername2   Runs on::Runs servername2

The best i have been able to do is to strip the server designation off, but since the value is still unique (a'la SYS_ID), it creates a "duplicate"...

What i need is to have one application automatically associate itself with all servers it finds:

provenirADMIN.exe   Runs on::Runs servername1 and servername2

I know this is possible, i am just having trouble finding the right instruction...

Anyone have an article handy?

[john_duchock]

"... you change it, you own it!"

Truer words have not been spoken.   I fully understand the implications of updating script level stuff in servicenow (upgrade exemption and what-not), so i will keep my efforts in our POV instance.   I will start down the path of the spkg effort and see about using it as a guide.

Thinking ahead, imagine what may have to be done when a software (application) is uninstalled or the server is decommissioned or otherwise no longer discovered.   I see the need for "cleaning" business rules, schedules and notifications to CI owners, etc...

...yeah, this may be a nose-bleed effort, but i have to give it a go  

[StephenHey]

If you're going to do nosebleed stuff, I would imagine you could probably:

Create a Business Service

Let discovery do its thing with applications

Modify discovery a bit to put 'application identifying' material in the application instance (if you have MySQL server used for MyCoolApp, put MyCoolApp information somewhere in the MySQL server record)

Create a many-to-many table with columns of Business Service, Application Class, and Application Identifying information

Create a Business rule that, upon update/insert of an application, queries against this MTOM table and creates relationships between the two.

Upon decommission of an application, business service, or CI, have a business rule deconstruct any of those relationships.   (Deconstruction happens automatically if you delete a CI, but if you're going to keep the CI, I don't think there's any business rules to do that.)

The method above would keep you much closer to OOB and your end users could then manage the MTOM table whenever a new application came online.

**EDIT** Also, depending on why you need to roll up your applications into one singular application, the Software Asset Management plugin might be helpful to you.   Upon discovery of an installed software, the system creates a Discovery Model.   These Discovery Models (can be multiple versions) then can be rolled up into one Software Model.   From there, you can perform Software Asset Management and count how many installs vs. licenses vs. entitlements you have.   Or, you can just see where this software is installed.