Bug 1661076

Summary:	Support node selector in metrics-server playbook
Product:	OpenShift Container Platform	Reporter:	Takayoshi Kimura <tkimura>
Component:	Monitoring	Assignee:	Frederic Branczyk <fbranczy>
Status:	CLOSED ERRATA	QA Contact:	Weinan Liu <weinliu>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	unspecified	CC:	anpicker, bjarvis, fshaikh, rekhan, surbania
Target Milestone:	---	Keywords:	Reopened
Target Release:	3.11.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:	2019-06-26 09:07:54 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:

Description Takayoshi Kimura 2018-12-20 03:49:21 UTC

Description of problem:

All non-app OpenShift components install come with node selector support but it seems that the metrics-server install is missing the node selector support.

Version-Release number of selected component (if applicable):

openshift v3.11.16

How reproducible:

Always

Steps to Reproduce:
1.
2.
3.

Actual results:

Users cannot specify node selector for metrics-server in Ansible, metrics-server goes to default node selector role=compute and consume app node resource.

Expected results:

Users can specify node selector for metrics-server in Ansible

Additional info:

Should metrics-server be installed by default?

Previously hawkular stack was optional but now we install Prometheus by default, so I feel it's better to install metrics-server by default along with Prometheus to align the changes.

Comment 1 Frederic Branczyk 2019-02-04 15:51:11 UTC

We intentionally didn't deploy metrics-server by default as in 3.11 we wanted to give users the choice. Starting with 4.0 we are shipping the prometheus-adapter by default (which offers the same API as the metrics-server, but backed by Prometheus).

Comment 3 Brian Jarvis 2019-05-03 12:47:46 UTC

The installer should not deploy metrics-server by default, but it should allow a user to specify a node selector for where the metrics-server pods should run.  

Many users separate worker nodes based on security rules, not allowing the user to specify where metrics-server pods run during the install forces users to manually modify the deployment.  This is a practice we should not be encouraging.

Comment 4 Frederic Branczyk 2019-05-03 15:56:25 UTC

Sorry I agree, this shouldn't have been closed. I think this is reasonable to add.

Comment 5 Frederic Branczyk 2019-05-10 15:07:38 UTC

Opened https://github.com/openshift/openshift-ansible/pull/11599 to fix this.

Comment 6 Frederic Branczyk 2019-05-10 17:59:38 UTC

PR merged

Comment 10 errata-xmlrpc 2019-06-26 09:07:54 UTC

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.

https://access.redhat.com/errata/RHBA-2019:1605