Hi all! Long time listener, first time caller. I was following the upgrade guide at Upgrading BEL using the lifecycle automation operator | Buoyant Enterprise for Linkerd yesterday and I came across something that didn’t work as expected, right at the very end. Specifically, after updating my BEL installations (I have five) using these instructions, and waiting for the control plane to go from Updating to UpToDate, I tried kubectl get dataplane -A and got the response No resources found rather than the expected output shown on the guide web page. Everything is working fine, though. Should I be alarmed that I apparently have no dataplanes?
There are two angles to this. One is that your workloads are meshed, which your statement, “Everything is working fine” would indicate. The second is whether your workloads are under the BEL Operator’s management. You can have meshed pods that aren’t managed.
Since you’re getting “No resources found” when issuing a kubectl get dataplane -A command, I’d guess that none of your namespaces are under Operator management. Have you created dataplanes for the namespaces you want the BEL Operator to manage? If not, follow this: Configuring BEL's lifecycle automation | Buoyant Enterprise for Linkerd
You will want to create a dataplane object for each namespace you wish to place under Operator management.
More information on lifecycle management can be found here: Lifecycle automation reference | Buoyant Enterprise for Linkerd
Hope that helps!
Tom
1 Like
Ah, nice! I see. I never configured Operator management at all; I just added the linkerd.io/inject annotation to the spec.template.metadata of each Deployment I wanted to mesh. I’ll look at installing the DataPlane manifests as well, and I’m sure that will resolve the No resources found message. Thanks!
My pleasure, Ben! Hope you enjoy the benefits of the Operator.