Good afternoon everyone,
More than a problem, I would like to report something that I consider like a bug.
Today it’s the first time that I use microk8s since I’m used to bigger environments, but I decided to give it a try for a development environment in order to have a reduced K8s Cluster for testing the applications before deploying them on the integration environment.
Well, I configured the basics as required and then I decided to include some custom configuration, mainly regarding Ingress and networking.
After checking, I thought into give a try to the modules and enable the Jaeger module on my microk8s-cluster with so bad luck that it didn’t recognize my already existing Ingress Controller and it automatically enabled a second module called ingress. Okay, no problem; I thought that I could just disable both modules and install the Jaeger Operator manually as I’m used to…
Problem is, that I’m used to put all my configuration related to Ingress in a custom created ingress namespace and when I disabled the ingress module: it deleted my ingress namespace!
Apparently the fact that the ingress module uses the same namespace gives the right to the module to delete it when you disable it.
From my point of view this is a huge bug:
Should I have to take care how do I name my namespaces just because some modules could enable others on which they depend? In case there is a namespace conflict and I disable a certain module all my namespace (created previously than that module enabled by itself) could be wiped?