Bug 1337595

Summary:	Anaconda's installclass should recognize RHEVH as a RHEL variant (and breaks EFI support)
Product:	Red Hat Enterprise Linux 7	Reporter:	Ryan Barry <rbarry>
Component:	anaconda	Assignee:	Samantha N. Bueno <sbueno>
Status:	CLOSED ERRATA	QA Contact:	Release Test Team <release-test-team-automation>
Severity:	high	Docs Contact:
Priority:	urgent
Version:	7.2	CC:	anaconda-maint-list, bcl, cshao, fdeutsch, jboggs, jneedle, jstodola, lmiksik, mbanas, rbarry, rvykydal, sbueno, weiwang, yaniwang, ycui
Target Milestone:	rc
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	anaconda-21.48.22.90-1	Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2016-11-03 23:24:48 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:	1329957, 1373387

Description Ryan Barry 2016-05-19 14:55:55 UTC

Description of problem:
If Product in /.buildinfo does not start with "Red Hat" (RHEV-H, for example), Anaconda will pick /fedora as the EFI path, as it has the highest priority.

Version-Release number of selected component (if applicable):
anaconda 21.48.22.56-1

How reproducible:
100%

Steps to Reproduce:
Change "Product" in /.buildinfo to start with anything other than "Red Hat"

Actual results:
"Fedora" has the highest priority

Expected results:
RHEV (and probably other variants) should be identified as RHEL, or have a way to override it other than changing the product. It's ok if this can be set with an additional field, to force selection to one of Fedora/CentOS/RHEL or other supported variants.

Comment 1 Fabian Deutsch 2016-05-19 14:58:10 UTC

This has a direct impact on EFI support, because the EFI loader location is determined by the install class, and the paths are different between RHEL and Fedora

Comment 2 Brian Lane 2016-05-19 17:51:46 UTC

I think it is much cleaner for products to identify themselves as "Red Hat ..." than it is for us to have to maintain a list of all the possible naming conventions that may come up.

Comment 3 Ryan Barry 2016-05-19 18:46:19 UTC

I agree, but "Red Hat Enterprise Virtualization Hypervisor" is a bit long for the Anaconda GTK interface.

The ability to override it (in upstream Anaconda) would also be ideal for oVirt, as those images are based off of CentOS, but tying the naming to "CentOS ..." or "Fedora ..." doesn't really work from an oVirt product perspective.

Comment 4 Fabian Deutsch 2016-05-23 12:36:07 UTC

I think we need to separate this bug, because there are two issues:

1. anaconda identifies RHEL/CentOS/Fedora based on the product title
2. anaconda screen exceeds display width it the title is to long (bug 1338793)

As there is a bug for 2., I'd suggest that we discuss 1. here.
It has to be considered that this bug becomes less relevant if bug 1338793 is getting fixed.

Because currently we need to choose a short title i.e. RHEV-H 3.6 Tech Preview, in which case anaconda falsely detects Fedora - which in turn breaks EFI (because the EFI path depends on the product).

Thoughts?

Comment 5 Fabian Deutsch 2016-06-16 09:50:19 UTC

FYI we are working on an install class:

https://github.com/evol262/anaconda/commit/e5e980b139344bf0a369e046320ab3ef1145308f

Once it is good we'll propose it in a PR in upstream.

Comment 6 Brian Lane 2016-06-16 16:15:23 UTC

I think the right way to do that would be to make, and maintain, your own rhev.py installclass and install it with a product.img or additional package when you create your live image.

Comment 7 Fabian Deutsch 2016-06-16 16:19:51 UTC

Right, we are actually rather at aiming to upstream our (very few) changes - the installclass is probably the only real change.

All other related changes are bugs which QE found in our flows, but which are in anaconda.

Do you have obejctions on including it in upstream?

Comment 8 Brian Lane 2016-06-16 17:41:53 UTC

(In reply to Fabian Deutsch from comment #7)
> Right, we are actually rather at aiming to upstream our (very few) changes -
> the installclass is probably the only real change.
> 
> All other related changes are bugs which QE found in our flows, but which
> are in anaconda.
> 
> Do you have obejctions on including it in upstream?

I do, yes, but that's probably a final decision for sbueno.

Comment 12 Samantha N. Bueno 2016-09-08 15:26:36 UTC

Patch is on review: https://github.com/rhinstaller/anaconda/pull/776

Setting this to POST.

I spoke to dougsland who said that the RHEV-H QE team can test for this specific fix, so RTT you can set this OtherQA.

Comment 13 Fabian Deutsch 2016-09-10 10:15:33 UTC

Ying, can you give qa_ack+ and cover testing?

Comment 30 errata-xmlrpc 2016-11-03 23:24:48 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://rhn.redhat.com/errata/RHEA-2016-2158.html