Bug 2230441

Summary:	cluster should be checked to be quorate and resources running after the role is finished green
Product:	Red Hat Enterprise Linux 9	Reporter:	michal novacek <mnovacek>
Component:	rhel-system-roles	Assignee:	Rich Megginson <rmeggins>
Status:	NEW ---	QA Contact:	CS System Management SST QE <rhel-cs-system-management-subsystem-qe>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	9.2	CC:	spetrosi
Target Milestone:	rc
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	role:ha_cluster
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:		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 michal novacek 2023-08-09 12:49:40 UTC

Description of problem:

As I understood running system roles is for most of the components only setting things up without verifying them.

Running rhel-system-roles.ha_cluster seems to be different because it makes no sense for most thing  to setting it up without cluster being running. My expectation would be that after running the role which ends green the cluster would be in a good state. 

Good state is when all the configure resources are in the desired (running?) state and the cluster is quorate.

This is especially painful as the role takes long time to run.


How reproducible: always


Steps to Reproduce:
1. run the role with correct configuration that is "wrong" for example incorrect fencing ips


Actual results: role finishing with green status but fencing not running 


Expected results: role finishing with green status only when cluster quorate and resources in desired (running?) state, red otherwise


Additional info:

This could be probably made optional to not break current workflow. 

It would probably useful to make resource/stonith/cluster quorate conditions separately selective.