Description of problem:
After added a windows worker to OCP 4.2 GA cluster, openshift-marketplace was trying to create pod on windows worker and failed, this should be avoided.
Version-Release number of selected component (if applicable):
OCP 4.2 GA
# oc get clusterversion
NAME VERSION AVAILABLE PROGRESSING SINCE STATUS
version 4.2.0 True False 27h Cluster version is 4.2.0
Steps to Reproduce:
1. Install OCP 4.2 GA and add windows worker
2. Delete redhat-operators pod in openshift-marketplace(maybe more than once)
# oc delete pod redhat-operators-5d878dc498-c74s4 -n openshift-marketplace
pod "redhat-operators-5d878dc498-c74s4" deleted
3. Check redhat-operators pod was trying to create in windows worker and failed
# oc get pod -n openshift-marketplace
NAME READY STATUS RESTARTS AGE
certified-operators-6b5c9d69b8-mftww 1/1 Running 0 27h
community-operators-8b5c5bf9-fvww6 1/1 Running 0 27h
marketplace-operator-5c846b89cb-c79qs 1/1 Running 0 27h
redhat-operators-547f57b5b8-vfddz 1/1 Running 0 27h
redhat-operators-5d878dc498-kxvjv 0/1 ContainerCreating 0 123m
4. Check redhat-operators deployment, NodeSelector is not specified
# oc edit deployment.apps/redhat-operators
redhat-operators pod was trying to create in windows worker and failed.
Avoid creating pod on windows worker for openshift-marketplace.
This can be accomplished by adding a NodeSelector to all builds:
Details please see: https://jira.coreos.com/browse/DEVEXP-341
*** This bug has been marked as a duplicate of bug 1774752 ***
1. the spec of nodeSelector of opsrc deployment:
2. delete the pods of opsrc the new pod won't be scheduled to the windows nodes.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.