Scheduled AutoDiscovery do not initiate modeling

Document ID : KB000047617
Last Modified Date : 14/02/2018
Show Technical Document Details

Issue: 

We have several scheduled AutoDiscoveries that run nightly. The discoveries complete, but modeling does not get triggered automatically, even though the "Discover and automatically model to SPECTRUM" option in the Discovery Console Configuration tab is set. The modeling works if we go back and manually select the "Model" button in the Discovery Tab after the discovery completes. Why is this happening?

 

Environment: 

This applies to all releases of Spectrum.

 

Cause:

When Spectrum runs a scheduled discovery, we use the permissions of the User model that created the discovery configuration to run the modeling. If this user model does not have the appropriate permission, or no longer exists in Spectrum, the modeling will not run. To help determine the problems enable the following debugs from the OneClick Web Server web page:

 

1. In a Web Browser, navigate to the OneClick Web page and go to the Administration Page

2. Select the "Debugging" link in the small gray header beneath the Tabs area

3. Select the "Web Server Debug Page (Runtime)" link on the left hand side of the page.

4. Enable the "AutoDiscovery", "Model Action Servlet", "User Access Privilege Manager", and "User Security" debugs

5. Go to the bottom of the page and ensure the debug level is set to "MAX" and click "Apply"

 

The debug output will be written to the $SPECROOT/tomcat/logs/stdout.log(Windows) or catalina.out (Solaris/LINUX) on the OneClick server.

 

Review the debugs to see if the following entries are logged:

(GAPM) - GAPM.getting licenses for user xxxxxxxxx mh null

 

(GAPM) - GAPM.***** User doesn't exist yet - NO privs

 

(ADISC) - ConfigMonitor.onJobCompleteEvent() - user (xxxxxxxxx) does not have permission to model.

 

In the example above, we see the user model does not exist in Spectrum and therefore the modeling failed to run.

 

 

Resolution:

When you create a AutoDiscovery Configuration in Spectrum, the discovery configuration is saved to the SSdb database as a model. To verify the User model that created the AutoDiscovery configuration, run a Locator search to find the "AdiscConfig" models. The "mdl_creat_athr" (0x11026) attribute shows the User model who created the AutoDiscovery Configuration.  You can check this by doing the following:

 

1. From the Locator tab in OneClick, run a Models search by "Model Type Name" = "AdiscConfig". This should provide a list of all the AutoDiscovery Configurations. 

2. Locate the scheduled discovery you are seeing the problem with. 

3. Select the AdiscConfig model and go to the Attributes Tab. 

4. Filter on creat

5. Select the "mdl_creat_athr" (0x11026) attribute to see which User model created the AutoDiscovery configuration. 

 

It should be something like USER@OC_HOST. 

 

Updating the "mdl_creat_athr' (0x11026) attribute to point to a valid user model in Spectrum resolves the issue, and the modeling automatically runs after the AutoDiscovery completes.

 

If the User model does exist, then check the User models permissions from the User Tab in OneClick and ensure they have the appropriate licensing, user roles, and privileges to allow them to create a model.

 

If everything appears to be correct and the modeling still fails, please open a case with CA Spectrum Support and provide copies of the sdtout.log/catalina.out containing the debugs for the support engineer to review.

 

 

Additional Information:

 

 

For more information about User permissions see the "OneClick Administration" guide - "User Administration in OneClick" section.

For more information about AutoDiscoveries see the "Modeling and Managing Your IT Infrastructure" guide - "Discovering and Modeling Your Network"