stack resource uuid link is broken

[Pavlo Shchelokovskyy]

Steps to reproduce:
- launch a stack by any means (CLI or via Dashboard)
- go to Orchestration -> Stacks -> Resources -> click on any resource
- the resource page is displaying a resource UUID as a link

Expected:
this link must lead to the actual resource page in the other OS component's space e.g. Compute

Actual:
the link is broken (404 Page not found)

There is clearly a spurious "/None" at the end of the link, but even without it the link would lead to the same page of the resource in Orchestration's space.

[Ana Krivokapić (akrivoka)]

Couldn't reproduce this bug using this simple template: https://raw.githubusercontent.com/openstack/heat-templates/master/hot/swift.yaml. The link on the resource page correctly link to the actual resource (in this case, a Swift container). Could you please provide an example template that can be used to reproduce this bug?

[Pavlo Shchelokovskyy (pshchelo)]

With master devstack (with masters of all components) I can reproduce this with this simplest template (using cirros 0.3.2 x64 image) - the link on my_instance Orchestration page is wrong as described above.

# This is a hello world HOT template just defining a single compute instance
heat_template_version: 2013-05-23

description: >
  HOT template that just defines single compute instance.

parameters:
  flavor:
    type: string
    description: Instance type for the instance to be created
    default: m1.nano
    constraints:
      - allowed_values: [m1.nano, m1.micro, m1.tiny, m1.small, m1.large]
        description: Value must be one of 'm1.nano', 'm1.micro', 'm1.tiny', 'm1.small' or 'm1.large'
  image:
    type: string
    description: name of the image to use for the instance
    default: cirros-0.3.2-x86_64-disk

resources:
  my_instance:
    type: OS::Nova::Server
    properties:
      image: { get_param: image }
      flavor: { get_param: flavor }

outputs:
  instance_ip:
    description: The IP address of the deployed instance
    value: { get_attr: [my_instance, first_address] }

[Amit Prakash Pandey (amitpp23)]

I have marked this bug as confirmed as I am getting the same error!

[Richard Jones (r1chardj0n3s)]

Can't reproduce this bug in latest Horizon using the above template (minor change: "cirros-0.3.2-x86_64-disk" has to be "cirros-0.3.2-x86_64-uec")

[Miguel Grinberg (miguelgrinberg)]

Reproduced on master devstack using template shown in comment #2 above.

[Miguel Grinberg (miguelgrinberg)]

For some reason the proposed fix I submitted earlier today for review did not get linked here. You can find it at https://review.openstack.org/#/c/116128/

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/116128
Committed: https://git.openstack.org/cgit/openstack/horizon/commit/?id=e8f5dc442c68d9ac206ae3a09bf5c18fe8125430
Submitter: Jenkins
Branch: master

commit e8f5dc442c68d9ac206ae3a09bf5c18fe8125430
Author: Miguel Grinberg <email address hidden>
Date: Thu Aug 21 23:00:02 2014 +0000

    show correct link to compute instances in stack resource overview tab

    There are two related fixes in the stacks resource overview page.
    First, any resource types for which horizon does not have a link for
    were getting a URL of "None" in the uuid field, and with this
    change the link is suppressed if it is None. The second part involves
    adding compute resources to the resource mappings, so that the uuid
    field for those links to the instance detail page.

    Change-Id: Ib8d51b49ac51092a3d7db6d889f4427167e5adb0
    Closes-Bug: 1322162