Red Hat Bugzilla – Bug 807639
References to "Frontend Realm" to be removed.
Last modified: 2012-12-13 14:49:44 EST
Description of problem:
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.create a cloud resource cluster for ec2 and do cluster mapping to ec2-us-east-1 and provider mapping to ec2-us-east-1: us-east-1a.
2.once the cloud resource cluster is created,click on Add mapping to cloud resource clusters and and again map to ec2-us-east-1.
"Frontend realm has already been taken" is displayed on UI.see attached screenshot.
Reference should be removed.
rpm -qa | grep aeolus
Created attachment 573317 [details]
ref to realm
The correct names for front and back end realms in Cloud Engine are "Cloud Resource Cluster" and "Provider Realm"
In the attached screenshot, the button labelled "Add mapping to cloud resource cluster" should read "Add mapping to Provider Realm"
"Cloud Resource Cluster test is mapped to Provider Resource Cluster:" should read "Cloud Resource Cluster test is mapped to Provider Realm:"
Created attachment 578711 [details]
Angus -- are you pointing out additional UI issues to be fixed as part of this bug?
I'm seeing a few different things here, and I'm having trouble what exactly is the reported bug, what is expected use of the app, and what is being deliberately added to expose a bug:
1) In the error message "frontend realm" should read "cloud resource cluster" -- this looks like a valid translation/overrides problem.
2) For this step: " do cluster mapping to ec2-us-east-1 and provider mapping to ec2-us-east-1: us-east-1a" -- are you only doing this to expose the duplication error? In practice it makes no sense to map to "ec2-us-east-1" _and_ to "ec2-us-east-1: us-east-1a". Mapping to "ec2-us-east-1" (i.e. a provider) says "match the provider as a whole; for this match we don't want to specify a provider realm, so leave that blank at launch". Mapping to "ec2-us-east-1: us-east-1a" says "match the ec2-us-east-1 provider and then, at launch, specify the provider realm "us-east-1a". In a real-world user situation, an admin would probably want to either map the provider (without realm) _or_ the provider realm -- the admin would _not_ want to map the provider (without realm) and, at the same time, map, separately, a realm for the _same_ provider.
3) Then, in step 2 you're making a third mapping -- to the same thing as the first mapping -- this is what triggers the 'name has been taken' error, since you're attempting to map the same provider twice.
4) Angus then points to another screenshot with two separate translation/overrides issues -- where we're using "Provider Resource Cluster" instead of "Provider Realm"
So, for the reported bug, I think the only actual bug reported is that we're using "Frontend Realm" instead of "Cloud Resource Cluster" -- although perhaps we should actually report a more obvious message to the effect that a provider (or provider realm) can only be mapped once to a Cloud Resource Cluster. Even better would be to, in addition, filter out the provider list to hide already-mapped providers and the realm list to hide already-mapped realms.
>> rpm -qa |grep aeolus
Considering Comment 4 above ... verfied that "Frontend Realm" is no longer mentioned. remapping the same realm gives the following message:
1 error prohibited this Cloud Resource Provider Realm Backend Target from being saved
Realm has already been taken
See attached screenshot.
Created attachment 615634 [details]
Realm Taken - no "frontend realm" mentioned