Goal:

The documentation needs to cover how to apply node-tolerations and node selectors at the namespace level for the namespace that will hold the QuayRegistry object.

Implementation:

See comment below: comment-15771546

Description:

This is an issue found when trying to binding Quay PODS to Openshift Infra Node, Now after adding taint to the infra node to prevent scheduling user workloads on, Quay Pods are still scheduled to OPC worker node, the reason is there's no tolerations defined for Quay PODS,  this issue was observed on Quay TNG Operator POD , Quay APP POD, Clair POD.

tolerations:
  - effect: NoSchedule 
    key: node-role.kubernetes.io/infra 
    operator: Exists 

Docs to bindling workloads to OPC Infra Nodes:

https://docs.openshift.com/container-platform/4.6/machine_management/creating-infrastructure-machinesets.html#assigning-machineset-resources-to-infra-nodes 

oc get node
NAME                                         STATUS   ROLES    AGE   VERSION
ip-10-0-128-71.us-east-2.compute.internal    Ready    worker   19h   v1.20.0+31b56ef
ip-10-0-137-36.us-east-2.compute.internal    Ready    master   19h   v1.20.0+31b56ef
ip-10-0-161-76.us-east-2.compute.internal    Ready    master   19h   v1.20.0+31b56ef
ip-10-0-180-92.us-east-2.compute.internal    Ready    worker   19h   v1.20.0+31b56ef
ip-10-0-206-228.us-east-2.compute.internal   Ready    worker   19h   v1.20.0+31b56ef
ip-10-0-213-56.us-east-2.compute.internal    Ready    master   19h   v1.20.0+31b56ef


oc adm taint nodes ip-10-0-137-36.us-east-2.compute.internal node-role.kubernetes.io/infra:NoSchedule
oc adm taint nodes ip-10-0-161-76.us-east-2.compute.internal node-role.kubernetes.io/infra:NoSchedule
oc adm taint nodes ip-10-0-213-56.us-east-2.compute.internal node-role.kubernetes.io/infra:NoSchedule

Index Image:

brew.registry.redhat.io/rh-osbs/iib:36350

Steps:

1. Login Openshift with OC login command
2. Add taint to all Openshift 3 Master nodes(see commands above), so those 3 Nodes come to OPC Infra Nodes
3. Create new OPC namespace and deploy Quay V3.4 TNG Operator
4. Check which OCP Node the TNG Operator is scheduled on
5. Deploy Quay with TNG Operator
6. Check which OCP Node the Quay App pod, config-editor POD, Clair are scheduled on

Expected Results:

All Quay PODs including TNG Operator POD are scheduled on OPC Infra Nodes.

Actual Results:

All Quay PODs including TNG Operator POD are scheduled on OPC Worker Nodes.