Bug 1942047

Summary:	[RFE][oVIrt] Add support for Machine guarantee memory field on Compute and Control Plain sections
Product:	OpenShift Container Platform	Reporter:	Guilherme Santos <gdeolive>
Component:	Installer	Assignee:	Evgeny Slutsky <eslutsky>
Installer sub component:	OpenShift on RHV	QA Contact:	Michael Burman <mburman>
Status:	CLOSED WONTFIX	Docs Contact:
Severity:	medium
Priority:	medium	CC:	eslutsky, gzaidman, mburman, suchaudh
Version:	4.7
Target Milestone:	---
Target Release:	---
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:	2021-11-04 12:47:16 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:	1970437
Bug Blocks:

Description Guilherme Santos 2021-03-23 14:15:30 UTC

Version:

$ openshift-install version
all

Platform:
ovirt/rhv

Please specify:
all

What happened?
when creating nodes, openshift set the vm memory value according with the user input. However, using api/sdk, rhv/ovirt doesn't automatically set guarantee_memory based on the memory input but it gets a default value from the blanket template, which is 1Gb. 
I don't think the nodes/vms running overcommiting the memory is very recommended and it can cause the nodes been unstable (specially in the installation, making it fail because takes longer than 30/20 min)


What did you expect to happen?
guarantee memory set according with user input memory

How to reproduce it (as minimally and precisely as possible)?


Anything else we need to know?
there is a ovirt/rhv RFE for it on rhv side: https://bugzilla.redhat.com/show_bug.cgi?id=1940529

Comment 1 Gal Zaidman 2021-03-24 10:42:17 UTC

Thanks Gui!

Douglas added you because it sounds like a fun fix on the installer

** Maybe terraform provider as well

Plus need to check that this is not happening on new machines, if that is the case then cluster-api-provider should be patched as well on a separate bug, Gui can you check that and open a new bug if needed?

Comment 2 Guilherme Santos 2021-03-24 14:18:43 UTC

On new machines do you mean like when we scale up and down the cluster?
If that is the case, yes. it does happen on new machines

Comment 4 Gal Zaidman 2021-03-30 13:35:07 UTC

due to capacity constraints we will be revisiting this bug in the upcoming sprint

Comment 10 Evgeny Slutsky 2021-05-03 14:36:54 UTC

@gal-  sure I'm working on it.

adding support for guranateed_memory for the terraform provider:

https://github.com/oVirt/terraform-provider-ovirt/pull/271

Comment 11 Evgeny Slutsky 2021-05-13 09:17:56 UTC

masters and workers are using the same rhcos template, so making this configuration part of the RHV template is problematic.
we can override the template configuration in VM level (same as setting the VM Memory)

Comment 12 Evgeny Slutsky 2021-06-10 13:30:54 UTC

need to implement in the VM  level. moving back to assigned

Comment 14 Gal Zaidman 2021-11-04 12:47:16 UTC

Closing this since it is not needed any more, the guaranteed memory issue only occurs in RHV 4.4.7 and we added a field to the machine object to be able to set it as a day2 op.
There is a KCS that explains all the changes across the RHV versions https://access.redhat.com/articles/6454811.

Please reopen if there is a need