Details
-
Bug
-
Resolution: Done
-
Undefined
-
None
-
None
-
None
-
False
-
None
-
False
-
Hypershift Sprint 232
-
0
-
0
-
0
Description
This ticket is linked with
https://issues.redhat.com/browse/SDA-8177
https://issues.redhat.com/browse/SDA-8178
As a summary, a base domain for a hosted cluster may already contain the "cluster-name".
But it seems that Hypershift also encodes it during some reconciliation step:
https://github.com/openshift/hypershift/blob/main/support/globalconfig/dns.go#L20
Then when using a DNS base domain like:
"rosa.lponce-prod-01.qtii.p3.openshiftapps.com"
we will have A records like:
"*.apps.lponce-prod-01.rosa.lponce-prod-01.qtii.p3.openshiftapps.com"
The expected behaviour would be that given a DNS base domain:
"rosa.lponce-prod-01.qtii.p3.openshiftapps.com"
The resulting wildcard for Ingress would be:
"*.apps.rosa.lponce-prod-01.qtii.p3.openshiftapps.com"
Note that trying to configure a specific IngressSpec for a hosted cluster didn't work for our case, as the wildcards records are not created.
