Bug 1252769

Created attachment 1061879 [details]
evm.log

Created attachment 1061880 [details]
production.log

I'm not sure if this is relevant, but here we go:

I have two providers in this appliance. RHEV 3.5 and OpenStack 6. There is a service catalog item to order resources in the RHEV environment.

This one always seems to work just fine.

If I create a service catalog item for OpenStack they work fine for some time too. It looks like the item breaks though, if something in the underlying OpenStack environment changes (like for example the network configuration). Also after some time (I don't really have been able to nail it down, but certainly after 24h) the item seems to be "broken".

So it might be related to changes in OSP which we update during inventory refresh and then break the catalog item.

It's only a theory of mine, but I thought it might help to mention it.

Regards,
Christian

Hi , 

I tried to recreate the issue at https://10.16.5.138. 
you can login as admin/smartvm . Under Service Catalogs , there is an item os_item . The template used to create the item was later deleted and so i see the error while editing that item . 

But not on clicking the item under Service Catalogs accordion.
Please check if this is the same error you were seeing .

Thanks, 
Shveta

Shveta, are you able to reproduce this on upstream??

The "undefined method 'base_class' for nil" error should be fixed by commits in https://bugzilla.redhat.com/show_bug.cgi?id=1250087

Not seeing this on upstream

Christian,

I can not reproduce this problem on upstream, 5.4.z, or on the appliance that you referenced.  Are there better instructions for reproducing it?  Does the catalog item need to be created with specific details?

Thanks,
Brandon

Closing since I can't reproduce the problem.  If there are better instructions to do so, please re-open with those.

Can comment #12 be used as a reproducer? 

I'm still trying to reproduce this in an internally accessible lab, but having problems with conflicting requests and availability.

Lucy - This seems similar to the work done in PR https://github.com/ManageIQ/manageiq/pull/2947 to allow catalog items to be modified after the template is deleted or orphaned from the provider.

Here it seems the selected Flavor is being deleted and I am not sure what would cause the blow up there.  Re-adding the flavor would not work because it will be associated with a new ID in the database.

Same basic failure as Bug 1236966

From the attached production.log:

[----] F, [2015-08-12T10:21:07.695808 #2593:12c3ea8] FATAL -- : Error caught: [NoMethodError] undefined method `base_class' for NilClass:Class
/var/www/miq/vmdb/app/models/miq_request_workflow.rb:1004:in `ci_to_hash_struct'
/var/www/miq/vmdb/app/models/miq_request_workflow.rb:999:in `add_target'
/var/www/miq/vmdb/app/models/miq_provision_cloud_workflow.rb:98:in `get_source_and_targets'
/var/www/miq/vmdb/app/models/miq_provision_virt_workflow.rb:639:in `get_source_vm'
/var/www/miq/vmdb/app/models/miq_provision_virt_workflow.rb:251:in `update_field_visibility'
/var/www/miq/vmdb/app/models/miq_provision_cloud_workflow.rb:72:in `update_field_visibility'
/var/www/miq/vmdb/app/models/miq_provision_virt_workflow.rb:40:in `initialize'
/var/www/miq/vmdb/app/controllers/application_controller/miq_request_methods.rb:748:in `new'
/var/www/miq/vmdb/app/controllers/application_controller/miq_request_methods.rb:748:in `prov_set_show_vars'
/var/www/miq/vmdb/app/controllers/catalog_controller.rb:1670:in `get_node_info'
/var/www/miq/vmdb/app/controllers/catalog_controller.rb:1758:in `replace_right_cell'
/var/www/miq/vmdb/app/controllers/catalog_controller.rb:313:in `tree_select'

Answered NEEDINFO via direct email

We have been unable to reproduce so I am closing this issue for now.  If the issue appears again please reopen the ticket with steps to reproduce.