Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1846671

Summary:	Highlight relantionship between defaultNodeSelector and pod's nodeSelector
Product:	OpenShift Container Platform	Reporter:	Olimp Bockowski <obockows>
Component:	Documentation	Assignee:	Michael Burke <mburke>
Status:	CLOSED CURRENTRELEASE	QA Contact:	RamaKasturi <knarra>
Severity:	medium	Docs Contact:	Vikram Goyal <vigoyal>
Priority:	medium
Version:	4.4	CC:	aos-bugs, jokerman, knarra, mburke
Target Milestone:	---
Target Release:	4.4.z
Hardware:	All
OS:	All
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2021-03-10 22:22:53 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	1846669
Bug Blocks:

Description Olimp Bockowski 2020-06-12 14:45:32 UTC

Document URL: 
https://docs.openshift.com/container-platform/4.4/nodes/scheduling/nodes-scheduler-default.html

Section Number and Name: 
"Understanding default scheduling"

Describe the issue: 
Readers could understand that having nodeSelector for Pod overwrites defaultNodeSelector that's not true

Suggestions for improvement: 
We should put some note that defaultNodeSelector & pod.spec.nodeSelector will be combined, e.g. Scheduler will try to find nodes with intersected (or unioned) selectors

Additional information: 
More details will be when #1846669 is updated

Comment 2 Michael Burke 2021-02-03 16:52:58 UTC

Doc PR: https://github.com/openshift/openshift-docs/pull/29102



I used this statement, which was added to the docs in a previous PR. 
"With cluster-wide node selectors, when you create a pod in that cluster, OpenShift Container Platform adds the default node selectors to the pod and schedules the pod on nodes with matching labels."
https://github.com/openshift/openshift-docs/pull/24865/files#diff-192126d49440cee046a761c16328f125c67ff2eabd68cb37ee8e025849adf632R10

Also used the wording in https://docs.openshift.com/container-platform/4.6/rest_api/config_apis/scheduler-config-openshift-io-v1.html#specification

defaultNodeSelector  This is applied to the pods created in all namespaces and creates an intersection with any existing nodeSelectors already set on a pod, additionally constraining that pod’s selector. Namespaces having project-wide node selectors won’t be impacted even if this field is set.

Comment 3 Michael Burke 2021-02-22 16:40:48 UTC

Xiaoli --

Can you please assign this BZ for QE review? Thank you very much.

Michael

Comment 4 Michael Burke 2021-03-01 16:45:28 UTC

Kasturi -- Is this an issue that you can review for QE?

Thank you
Michael

Comment 5 RamaKasturi 2021-03-04 09:04:08 UTC

Hi Michael,

   I have reviewed the PR and add /lgtm label. Let me know if the bug here can be moved to verified state now, i can do that, thanks !! Since i am the QE contact and did what all is needed for this bug, clearing need infos. Thank you !!

Comment 6 Michael Burke 2021-03-04 13:23:15 UTC

Rama -- Yes, when you approve of the changes in the PR, you can move the BZ to Verified. 

Thank you for the review!

Michael

Comment 7 Michael Burke 2021-03-04 20:14:10 UTC

Changes are live: https://docs.openshift.com/container-platform/4.7/nodes/scheduling/nodes-scheduler-node-selectors.html#nodes-scheduler-node-selectors-pod_nodes-scheduler-node-selectors

Comment 8 Olimp Bockowski 2021-03-08 07:43:46 UTC

super, thanks for your work