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
{{ message }}
This repository has been archived by the owner on May 24, 2023. It is now read-only.
We recently switched to 0.3.0, and with that get an error that own namespace is not supported (did not test 0.2.0). It seems that currently, we can only let the operator group be effective cluster-wide, which we do not want to do. We are running a multi-tenant environment (OpenShift 4.7), and want the operator to only pick up NginxIngressController instances from its own namespace. It would be great if the previous behavior could be restored.
The text was updated successfully, but these errors were encountered:
I'm trying to understand your setup. Do you have multiple nginx-ingress-operator running in the same cluster, but in different namespaces?
IIRC the operator was always watching all the namespaces, no matter where it was installed and the new one behaves in the same way, but it can only be installed in the dedicated namespace.
We deploy the nginx ingress controller operator once, using OLM. We then roll out a 3-replica nginx ingress controller instance by applying a corresponding NginxIngressController custom resource manifest in a designated namespace, to which only we as cluster operators have access. This nginx ingress controller instance will watch cluster wide.
We want to prevent everyone else in the cluster from rolling out additional nginx ingress controller instances. This should be achieved by letting the nginx ingress controller operator deployed via OLM watch for NginxIngressController custom resources exclusively in the designated namespace.
Thanks for the explanation and feedback.
From what I can see in the code, it seems like that the Operator was always watching all namespaces, the only difference is that now you can't select the namespace where you want to install it.
With that said, we'll be working on (properly) supporting the targetNamespaces option soon, stay tuned 🙂
So far, we have been using version 0.1.0, and in the
OperatorGroup
'stargetNamespaces
, we set the namespace where the operator is deployed:We recently switched to 0.3.0, and with that get an error that own namespace is not supported (did not test 0.2.0). It seems that currently, we can only let the operator group be effective cluster-wide, which we do not want to do. We are running a multi-tenant environment (OpenShift 4.7), and want the operator to only pick up
NginxIngressController
instances from its own namespace. It would be great if the previous behavior could be restored.The text was updated successfully, but these errors were encountered: