Bug 1818577 (OCPRHV-61-4.5)

Summary:	[rhv] installer created template has multiple shortcomings, esp disk size
Product:	OpenShift Container Platform	Reporter:	daniel <dmoessne>
Component:	Installer	Assignee:	Roy Golan <rgolan>
Installer sub component:	OpenShift on RHV	QA Contact:	Jan Zmeskal <jzmeskal>
Status:	CLOSED DUPLICATE	Docs Contact:
Severity:	high
Priority:	high	CC:	erich, jcall, lsvaty, michal.skrivanek, mkalinin, mmuench, mtessun, pelauter, rdlugyhe, rgolan, sbonazzo, trees, vbhatia
Version:	4.4	Flags:	trees: needinfo- trees: needinfo-
Target Milestone:	---
Target Release:	4.5.0
Hardware:	Unspecified
OS:	Unspecified
URL:	https://issues.redhat.com/browse/OCPRHV-61
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:
Clones:	OCPRHV-61-4.4.z (view as bug list)		Environment:
Last Closed:	2020-05-19 11:32:04 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:
Bug Blocks:	1823809

Description daniel 2020-03-29 13:05:04 UTC

Description of problem:

openshift installer ipi for rhv has some shortcomings, i.e.:

os type should rather be rhel 8 than "Other OS"
optimized for server rather than desktop 
--> those have been already addressed by bz 1814042

and in my eyes the worst, the generated disk size is just 16GiB which is way too small and already leads to alters right after install that the disk on the masters is quite full. 


Version-Release number of the following components:
 rhv 4.3.7.2
- oc version 
Client Version: 4.4.0-rc.4
Server Version: 4.4.0-rc.4
Kubernetes Version: v1.17.1

How reproducible:

Steps to Reproduce:
1. install OCP 4 on RHV as outlined by https://access.redhat.com/articles/4903411
2. after installation check events/alerts

Actual results:
alerts firing complaining about high disk usage on masters

Expected results:
disk has a sensible size not causing issues right after installation 

Additional info:
The docs provide a procedure how to customize the template and explicitly mention how the disk size could be increased, however I think the installer should initially create a disk size that is not right after install causing alarms but has a sensible size.
Looking through the docs there are several different sizes depending on the underlying infra.
I very much tend to the recommendations of vmware: https://docs.openshift.com/container-platform/4.3/installing/installing_vsphere/installing-vsphere.html#minimum-resource-requirements_installing-vsphere which would be 120GiB disk size as we could do it thin provisioned to save space and as this template is used for the workers as well with custom load running on it the small initial disk size is really annoying and is immediately leading to problems.

Comment 7 Michal Skrivanek 2020-04-08 12:34:10 UTC

(In reply to daniel from comment #0)
> openshift installer ipi for rhv has some shortcomings, i.e.:
> 
> os type should rather be rhel 8 than "Other OS"

RHCOS type - rhcos_64 - since 4.3.9

> optimized for server rather than desktop 

High Performance VM profile would likely be a much better fit

Comment 8 Michal Skrivanek 2020-04-14 13:30:01 UTC

*** Bug 1814042 has been marked as a duplicate of this bug. ***

Comment 9 Michal Skrivanek 2020-04-14 13:33:04 UTC

defaul disk size would definitely need a 4.4.z backport, others may wait. Pending on patch readiness...

Comment 10 Michal Skrivanek 2020-04-14 13:53:28 UTC

OS type is tracked in bug 1821638

Comment 11 Marina Kalinin 2020-05-01 18:01:58 UTC

Hi all,

We have created this KCS to track customers that hit this bug in their deployment and don't read documentation.
https://access.redhat.com/solutions/5043561

Anything to add to it? 
Maybe any troubleshooting steps or more detailed message?

Thank you.

Comment 21 John Call 2020-05-06 18:20:25 UTC

(In reply to Rolfe Dlugy-Hegwer from comment #20)
> - We can't assume every use case is a production cluster.
> - We must consider resource-constrained use cases.
I agree that non-production use cases are important.  I personally have to setup many proof-of-concept, lab, test, and demo clusters.  It would benefit me to know what the "minimum" required resources are for those environments.  But I still believe that 120GB should be the most visable value.  If we show 32GiB we will see too many support tickets, and too many emails to Solution Architects asking why the cluster stopped working.


> --
> https://docs.openshift.com/container-platform/4.4/installing/installing_rhv/
> installing-rhv-creating-custom-vm.html#installing-rhv-using-ansible-
> playbook_installing-rhv-creating-custom-vm clearly states "Specify 32GiB,
> the default value, or more. 120GiB is the minimum for production
> environments."
May I suggest that this sentence be changed from "32GiB, the default value..." to "32GiB, the minimum value for non-production clusters, or more."

Comment 29 Sandro Bonazzola 2020-05-18 16:11:29 UTC

Also tracked in Jira at: https://issues.redhat.com/browse/OCPRHV-61

Comment 30 Jan Zmeskal 2020-05-19 11:32:04 UTC

This has actually been taken care of in https://bugzilla.redhat.com/show_bug.cgi?id=1820300

*** This bug has been marked as a duplicate of bug 1820300 ***