Closed tvandinther closed 3 weeks ago
This may be fixed upstream if the dynamic client is amended to find cluster-scoped resources even when a namespace is set as per this issue: https://github.com/kubernetes/kubernetes/issues/111248
It also seems that this stale issue covers the same problem: https://github.com/argoproj/argo-events/issues/3132
This may be fixed upstream if the dynamic client is amended to find cluster-scoped resources even when a namespace is set as per this issue: kubernetes/kubernetes#111248
It also seems that this stale issue covers the same problem: #3132
Would you like to submit a PR to add those (maybe not all but often used) cluster scope resources?
If you're happy with that approach I can submit the PR.
Describe the bug When using the Kubernetes object trigger to delete a cluster-scoped resource (such as a persistent volume), the message
object not found, nothing to delete...
is logged.To Reproduce Delete any cluster-scoped resource (other than
namespace
ornode
) using the Kubernetes object trigger.Expected behavior The resource should be identified as cluster-scoped and found and deleted appropriately.
Environment (please complete the following information):
Additional context It is possible this issue affects operations other than
Delete
. I noticed that in this part of the code, only namespaces and nodes are listed as cluster-scoped resources. Following a static analysis of the code, I presume the namespace is set to the sensor namespace and subsequently not found.I'd expect that this issue is fixed by making the list exhaustive (for standard resources), or a new mechanism used to determine cluster-scoped resources so that cluster-scoped CRs can also be identified.
Message from the maintainers:
If you wish to see this enhancement implemented please add a 👍 reaction to this issue! We often sort issues this way to know what to prioritize.