Bug 1738456

Summary: [DOCS] master would become schedulable in all UPI install
Product: OpenShift Container Platform Reporter: Johnny Liu <jialiu>
Component: DocumentationAssignee: Kathryn Alexander <kalexand>
Status: CLOSED CURRENTRELEASE QA Contact: Johnny Liu <jialiu>
Severity: urgent Docs Contact: Vikram Goyal <vigoyal>
Priority: high    
Version: 4.2.0CC: aos-bugs, dhansen, jiajliu, jokerman, wking
Target Milestone: ---   
Target Release: 4.2.z   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2019-10-17 15:42:15 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:

Description Johnny Liu 2019-08-07 08:48:33 UTC
Document URL: 
https://docs.openshift.com/container-platform/4.1/installing/installing_bare_metal/installing-bare-metal.html#installation-bare-metal-config-yaml_installing-bare-metal
https://docs.openshift.com/container-platform/4.1/installing/installing_aws_user_infra/installing-aws-user-infra.html#installation-generate-aws-user-infra_installing-aws-user-infra
https://docs.openshift.com/container-platform/4.1/installing/installing_vsphere/installing-vsphere.html#installation-vsphere-config-yaml_installing-vsphere

Section Number and Name: 
Sample install-config.yaml file

Describe the issue: 
In our existing doc, we are telling user to set worker replica to 0, while since https://github.com/openshift/installer/blob/master/pkg/asset/manifests/scheduler.go#L64-L73 is merged, that would make all masters become schedulable. That means user's pods maybe scheduled to any master node. I do not think this is what user expected.
Per https://github.com/openshift/installer/pull/2004, there are long discussion about this new change, what we should tell user to do in our 4.2 official doc? What is our finial recommendation. This is a big change, at least a massive change is required in doc.

Suggestions for improvement: 

Additional information:

Comment 4 Kathryn Alexander 2019-10-02 19:47:43 UTC
PR's here: https://github.com/openshift/openshift-docs/pull/17016

I'm asking Trevor to take a look before I put it on QE.

Comment 5 Kathryn Alexander 2019-10-04 14:30:43 UTC
@Jianlin, will you PTAL? https://github.com/openshift/openshift-docs/pull/17016

Comment 6 Johnny Liu 2019-10-11 06:47:43 UTC
upi on aws looks good to me.
Where is upi on vsphere, and upi on baremetal PR?

Comment 9 W. Trevor King 2019-10-14 22:40:24 UTC
*** Bug 1755073 has been marked as a duplicate of this bug. ***

Comment 12 Kathryn Alexander 2019-10-15 13:20:07 UTC
I updated the PR to add the steps to generate the manifest files and modify the cluster-scheduler-02-config.yml file to vsphere and bare metal.

Comment 13 Kathryn Alexander 2019-10-15 13:20:29 UTC
@Jianlin, will you please take another look?

Comment 14 Johnny Liu 2019-10-16 02:09:33 UTC
The PR for aws/gcp looks good to me.

Waiting for new PR for vsphere and baremetal.

Comment 15 Kathryn Alexander 2019-10-16 02:21:26 UTC
Because of the change I had to make to cover vSphere and bare metal, I added them to this PR. (eg https://bz1738456--ocpdocs.netlify.com/openshift-enterprise/latest/installing/installing_vsphere/installing-vsphere.html#installation-user-infra-generate-k8s-manifest-ignition_installing-vsphere)

Comment 16 Johnny Liu 2019-10-16 02:34:33 UTC
Ah, I lost your new change. LGTM.

Comment 17 Kathryn Alexander 2019-10-16 12:23:38 UTC
Thank you! My apologies for the confusion. I've merged the change and am waiting for it to go live.