Description of problem: The downloads pods deploy on woker nodes by the console operators. But some users don't like that, so please improve it so that it can be set to master fixed. Version-Release number of selected component (if applicable): OCP4.6 How reproducible: After installed cluster. $ oc get pod -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES console-857895b985-4tbzr 1/1 Running 0 7d18h 10.128.0.4 ip-10-0-206-223.ap-northeast-1.compute.internal <none> <none> console-857895b985-swv87 1/1 Running 0 7d18h 10.129.0.21 ip-10-0-158-151.ap-northeast-1.compute.internal <none> <none> downloads-5b94cdbbf4-69cmk 1/1 Running 0 7d18h 10.131.0.11 ip-10-0-184-184.ap-northeast-1.compute.internal <none> <none> downloads-5b94cdbbf4-g77x9 1/1 Running 0 7d18h 10.128.2.19 ip-10-0-216-121.ap-northeast-1.compute.internal <none> <none> ]$ oc get node NAME STATUS ROLES AGE VERSION ip-10-0-135-95.ap-northeast-1.compute.internal Ready worker 20d v1.19.0+d670f74 ip-10-0-158-151.ap-northeast-1.compute.internal Ready master 20d v1.19.0+d670f74 ip-10-0-180-101.ap-northeast-1.compute.internal Ready master 20d v1.19.0+d670f74 ip-10-0-184-184.ap-northeast-1.compute.internal Ready worker 20d v1.19.0+d670f74 ip-10-0-206-223.ap-northeast-1.compute.internal Ready master 20d v1.19.0+d670f74 ip-10-0-216-121.ap-northeast-1.compute.internal Ready worker 20d v1.19.0+d670f74 Actual results: The downloads pods deploy on woker nodes by the console operators. Expected results: They should be on controle plane nodes. Additional info: The console pods deploy on mater nodes by the NodeSelector. https://github.com/openshift/console-operator/blob/5c37bd88347626255fb043a7b23cda71e90de3f3/pkg/console/subresource/deployment/deployment.go#L118 ``` Template: corev1.PodTemplateSpec{ ObjectMeta: metav1.ObjectMeta{ Name: api.OpenShiftConsoleName, Labels: labels, Annotations: podAnnotations, }, Spec: corev1.PodSpec{ ServiceAccountName: "console", // we want to deploy on master nodes NodeSelector: map[string]string{ // empty string is correct "node-role.kubernetes.io/master": "", }, ```
Closing since moving to a jira story - https://issues.redhat.com/browse/CONSOLE-2903