ASB support multi registry with the same name currently

[zihantang-rh]

This form is for bug reports and feature requests. Major features will go through a spec process.

What happened:
In v3.7, if config broker-config with multi registry with the same name , the asb pod will fail with errror
ERROR: Failed to read config file registry name must be unique
But in v3.9 the asb pod will start successfully. And successfully loaded apbs from the last registry.
Is it the expected feature ? @jwmatthews

openshift v3.9.0-0.19.0
kubernetes v1.9.0-beta1
ASB: brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/openshift3/ose-ansible-service-broker:v3.9 [1.1.4]

What you expected to happen:

How to reproduce it:

1. Config broker-config with:

• type: dockerhub
  name: test
  url: https://registry.hub.docker.com
  org: ansibleplaybookbundle
  tag: latest
  white_list: [.*-apb$]
• type: dockerhub
  name: test
  url: https://registry.hub.docker.com
  org: zjianbjz
  tag: latest
  white_list: [.*-apb$]

2. Restart broker.
   Logs:
   [2018-01-16T09:19:30.029Z] [DEBUG] - APBs passing white/blacklist filter:
   [2018-01-16T09:19:30.029Z] [DEBUG] - -> zjianbjz/fail-apb
   [2018-01-16T09:19:30.029Z] [DEBUG] - -> zjianbjz/hello-world-db-apb
   [2018-01-16T09:19:30.029Z] [DEBUG] - -> zjianbjz/mediawiki-apb
   [2018-01-16T09:19:30.029Z] [DEBUG] - -> zjianbjz/postgresql-apb
   [2018-01-16T09:19:30.029Z] [DEBUG] - -> zjianbjz/hello-test-apb
   [2018-01-16T09:19:30.117Z] [DEBUG] - Registry::imageToSpec
   [2018-01-16T09:19:30.981Z] [NOTICE] - Broker successfully bootstrapped

[jmrodri]

This looks related to the recent config refactor we did.

[rthallisey]

Since the org field is different, we will likely get a different set of images. So this is the functionality we want.

[shawn-hurley]

@rthallisey I think that is the issue. The FQName only takes into account of the registry name and the name defined from the spec yaml. So dh-mediawiki123-apb, where registry name is dh and apb name is mediawiki123-apb. I think that this is a regression bug and as @jmrodri has mentioned is due to the config refactor.

[shawn-hurley]

Create a bugzilla for this issue:
https://bugzilla.redhat.com/show_bug.cgi?id=1535182

Will keep both open to track this regression.