Add cluster-identity#429
Conversation
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
mjudeikis
force-pushed
the
identity.add
branch
4 times, most recently
from
f7984df to
2456887
Compare
There was a problem hiding this comment.
So we are creating one namespace per author/"consumer" to store the cluster objects?
I wonder if it wouldn't be safer to make the namespace deterministic instead of using generateName, e.g. kube-clusters-{identity-hash}; that would prevent things like one identity "owning" multiple namespaces.
Edit: I wouldn't change the namespace handling in this PR. We can adjust that later; just came to mind while reading the code.
| // +kubebuilder:validation:Required | ||
| TemplateRef APIServiceExportTemplateRef `json:"templateRef"` | ||
| // ClusterIdentity contains information that uniquely identifies the cluster. | ||
| // When doing dry run, we expect the client to fill this field in (or it will be taked from local cluster where context is available). |
There was a problem hiding this comment.
| // When doing dry run, we expect the client to fill this field in (or it will be taked from local cluster where context is available). | |
| // When doing dry run, we expect the client to fill this field in (or it will be taken from local cluster where context is available). |
| identity string | ||
| author string |
There was a problem hiding this comment.
Author is the user and identity is the cluster identity?
Signed-off-by: Mangirdas Judeikis <mangirdas@judeikis.lt> On-behalf-of: @SAP mangirdas.judeikis@sap.com
mjudeikis
force-pushed
the
identity.add
branch
from
2456887 to
204e08a
Compare
I think we can make this with this change too. Its kinda easy and either way - small breaking change. |
mjudeikis
force-pushed
the
identity.add
branch
from
4e24ea0 to
f2874d4
Compare
|
made the ns static base36 change. same as kcp uses for logical clusters |
2 participants
Summary
Re-add cluster-identity as a binding request. This would allow pre-provisioning of right identity namespaces based on cluster identity, rather than user identity (as it is now).
This is a temporary solution until we can sort this out in a more sustainable way. Its bit hacky.
What Type of PR Is This?
/kind regression
/kind bug
Related Issue(s)
Fixes #
Release Notes