-
Notifications
You must be signed in to change notification settings - Fork 63
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add provider form fields revisit for DDF conversion #383
Comments
If I had to guess its because it is in its own repo and isn't only added as a child manager. Doesn't make it a good reason and I'm open to changing it 😄 If we don't expect to ever have kubevirt as a standalone manager without tying it to a kube/ocp (which seems reasonable) we can change that provider to delegate its auth to a parent_manager. |
Yeah, kubevirt is custom resources accessed over same apiserver, From what i remember, the way containers child managers update child Endpoints/Auth records in rails hooks when parent manager changes (or something like that) was a horrible kludge that bit us in subtle ways :-( IIRC @masayag tried to apply this pattern to kubevirt and had issues with it (?) |
If that's okay, I'd like to get rid of that token field for kubevirt, it would not need any special handling on the backend after converting the form to DDF. |
KubeVirt is an add-on for kubernetes which requires different roles to fully-operate than the roles that are needed for the container manager. That's the main reason why we added a specific token for the virtualization endpoint. |
@skateman that sounds reasonable to me. So eventually the 'virtualization' sub-tab will contain only the token field and validate button ? |
Yes, as that's the only editable field there... |
@agrare there are similar delegation problems here as we had in configuration management. |
Well that's "fun" |
Closing since the question was answered |
I'm trying to convert the add-provider form for containers to DDF and there are some inconsistencies. All the endpoints use the same token for authentication, this is solved by copying the default record. This would not be a problem, it is all sorted out on the model-level. However, the virtualization requires a separate token and on my setup I was able to use the same token as for the default endpoint.
Is the separate token option really necessary? It would help me a lot if it would just use the same mechanism as any other endpoint.
Parent issue: ManageIQ/manageiq#18818
The text was updated successfully, but these errors were encountered: