Bug 975520

Summary:	content availability based on arches is currently too tolerant
Product:	[Community] Candlepin	Reporter:	John Sefler <jsefler>
Component:	candlepin	Assignee:	Adrian Likins <alikins>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Katello QA List <katello-qa-list>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	0.9	CC:	dgregor, jesusr
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2013-10-01 13:50:04 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:	840995

Description John Sefler 2013-06-18 16:51:34 UTC

Description of problem:

ISSUE 1:
The new "arches" attribute on a content set was designed to satisfy RFE bug 706187.  Currently the implemented design is letting content through to the entitlement on a system whose arch is not on the content set arches list.

For example, here's a log statement from an automated test that is failing...

> java.lang.AssertionError: As expected, contentNamespace label 'content-label-64' defined for arches 'i386,i686' is NOT included in the entitlement after subscribing to 'My Subscription providing Arch-Based Content' on a 'x86_64' system. expected:<true> but was:<false>

What's happening is that Candlepin is trying to be too smart.  Although the arches for this content set have been set to "i386,i686" and the system's arch is "x86_64", Candlepin is letting this content set through because Candlepin is smart enough to know that i386,i686 packages are runnable on an x86_64 system.  However, discussions with dgregor from release engineering indicate that Candlepin should honor the arches defined on a content set more literally.  This example content set should only be available on a system whose arch is "i386" or "i686".  If release engineering wanted this content set available on an x86_64 system too, then the arches would be defined by release engineering as "i386,i686,x86_64".


ISSUE 2:
The same story applies to "ppc".  A content set defined with arches="ppc" should not be available via the candlepin entitlement on a system whose arch is "ppc64".  If this is what release engineering wanted, then they will set the content set arches="ppc,ppc64".


ISSUE 3:
Another request that is not yet implemented is that Candlepin should tolerate "x86" on the content set "arches" attribute as an alias for "i386,i486,i586,i686".  Release-engineering would like to specify an arches value of "x86" and then this content would be available on system's whose arch is either i386 or i486 or i586 or i686 (but not x86_64).




Version-Release number of selected component (if applicable):
[root@jsefler-f14-candlepin ~]# cd candlepin
[root@jsefler-f14-candlepin candlepin]# git show-ref | grep master
2c62529c399083cd29a50371c51478053c1c0919 refs/heads/master

Comment 1 John Sefler 2013-06-19 14:06:19 UTC

ISSUE 4:
Another request that is not yet implemented is that Candlepin should tolerate "noarch" on the content set "arches" attribute to behave exactly like the "ALL" alias which is already implemented.  Hence if "noarch" or "ALL" is set within the content set "arches", then it will be included in the entitlement and available on any system.

Comment 2 Adrian Likins 2013-07-01 19:41:48 UTC

commit ef7bd57af96d2a4b11901a2887be88213788d339
Author: Adrian Likins <alikins>
Date:   Tue Jun 18 16:02:15 2013 -0400

    make Content arch compares more specific
    
    Content for example, for 'i686' now only
    matches 'i686' machines (no longer matches
    as compatible with 'x86_64')
    
    Add support for Content arch of 'x86' that
    maps to i386,i486,i586,i686

Comment 3 John Sefler 2013-07-18 22:10:00 UTC

Verified Version...
[root@jsefler-f14-candlepin candlepin]# git show-ref | head -1
57ac1f8da3ad1abe463a5ec6179813bf707daa35 refs/heads/master

[root@jsefler-5 ~]# subscription-manager version
server type: Red Hat Subscription Management
subscription management server: 0.8.16-1
subscription-manager: 1.8.13-1.el5
python-rhsm: 1.8.13-1.el5


The automated test coverage that originally detected this bugzilla is now passing.
VerifyContentSetsEntitledFromSubscriptionPoolSatisfyTheSystemArch_Test(SubscriptionPool pool)
VerifyArchRestrictedContentSetsEntitledAfterSubscribeAllSatisfiesTheSystemArch_Test()

1. Content sets with an "x86" arches value are now being granted to a consumer whose system arch is among "i386","i486","i586","i686"

2. Content sets with an arches value containing one or more  "i386,i486,i586,i686,x86" are no longer being granted to a consumer whose system arch is "x86_64".

3. Content sets with a "noarch" arches value are now granted to a consumer no mater what their system arch value is.


Moving to VERIFIED.