You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The operator rewrote the provider kubeconfig host to /clusters/<providerWorkspace>
(a path). But the admin-portal provider kubeconfig is already scoped to the
provider workspace by cluster ID, and front proxies that route by cluster ID 404
on the path form — so bootstrap failed with "the server could not find the
requested resource" when creating CRDs.
- RetargetHostToWorkspace now leaves an already-scoped (non-root) host untouched;
it only retargets a root-scoped / unscoped host (the dev admin-kubeconfig case).
- providerWorkspace is now optional: when unset, the operator discovers the path
from the workspace's kcp.io/path annotation (LogicalCluster). It's only needed
for stamping spec.export.path on the APIExportEndpointSlice.
- CRD: providerWorkspace +optional (regenerated). Chart defaults it to "" and the
CR omits it when empty.
So a workspace-scoped provider kubeconfig needs no providerWorkspace and is used
as-is.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
0 commit comments