-
Bug
-
Resolution: Unresolved
-
Major
-
None
-
4.18
-
None
-
Quality / Stability / Reliability
-
False
-
None
-
None
-
Important
-
None
-
None
-
None
-
None
-
None
-
Done
-
Known Issue
-
-
None
-
None
-
None
-
None
Description of problem:
When deploying a cluster with nodes of multiple architectures (e.g., both arm64 and amd64), the nmstate-handler DaemonSets are scheduled only on nodes with amd64 architecture. This is due to the presence of the nodeSelector in the DaemonSet configuration, which restricts pod scheduling to amd64 nodes only. kubernetes.io/arch=amd64 kubernetes.io/os=linux As a result, nmstate-handler is not deployed on nodes with other architectures (e.g., arm64), limiting the functionality of the nmstate operator in heterogeneous clusters.
Version-Release number of selected component (if applicable):
4.18
How reproducible:
100%
Steps to Reproduce:
1. Create a cluster with both amd64 and arm64 architecture nodes. 2. Install the nmstate operator and create an nmstate custom resource. 3. Observe that the nmstate-handler DaemonSet is deployed only on amd64 nodes, and no pods are scheduled on arm64 nodes due to the restrictive nodeSelector.
Actual results:
nmstate-handler DaemonSet is deployed only on amd64 nodes, and no pods are scheduled on arm64 nodes.
Expected results:
nmstate-handler DaemonSet should run on each nodes.
Additional info: