Bug 767598

Summary:	Pushing RHEL images to EC2 using system template containing on premise repos fails
Product:	[Retired] CloudForms Cloud Engine	Reporter:	chris alfonso <calfonso>
Component:	oz	Assignee:	Ian McLeod <imcleod>
Status:	CLOSED CURRENTRELEASE	QA Contact:	James Laska <jlaska>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	1.0.0	CC:	brad, hbrock, jturner, whayutin
Target Milestone:	rc
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	oz-0.7.9-7.el6	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2012-01-16 20:38:08 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:

Description chris alfonso 2011-12-14 13:26:23 UTC

Description of problem:
When pushing a RHEL image to EC2 from an on premise CE, if the system template used to build the image contains repo urls that are internal hostnames, the push will fail due to the non-routability of the hostname.

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


How reproducible:
Set up a CloudForms, get a system template from SE and make sure it has internal only accessible repositories in the template.  Use that template to build a RHEL image and then push it to EC2.  The push will fail because it  can't route to the repository hosts.

Actual results:
The push of the image fails.

Expected results:
The push of the image succeeds and any packages specified in the template are installed from the internal repo.

Additional info:
A patch has already been created, posted.  Awaiting review.

Comment 1 jrd 2011-12-14 15:04:47 UTC

I thought we said we weren't supporting the pushing of rhel images to ec2?

Perhaps this is really a build issue, not a push issue?  For build time, there's supposed to be an ssh tunnel set up to deal with "reaching back" to get privately hosted packages.  That's likely the problem.

Comment 2 chris alfonso 2011-12-14 16:06:17 UTC

The image customization happens during the push. The patch has been posted for review.

Comment 3 wes hayutin 2011-12-15 15:00:52 UTC

we need a git hash in the bug so eggs can pull it in to the build

Comment 4 wes hayutin 2012-01-03 17:43:17 UTC

adding ce-sprint-next bugs to ce-sprint

Comment 5 chris alfonso 2012-01-04 14:27:01 UTC

patches pushed to oz repos:

51ae6391868f0571515b4cc6dc2cb3e75d56605a
4a6501508d9c757a20607b7c1054b86fe649928cq
6d6b58e27839a9eb97d9da1d57685573503bf843
da3e28bf8b34c0a2035de383cc0c5535ceb9a533

Comment 6 Ian McLeod 2012-01-05 17:45:36 UTC

This is brewed internally as 0.7.9-6 (which includes all of the commits referenced above and is, in fact, current with git master as of today, January 5 2012.)

Switching to ON_QA

Comment 7 wes hayutin 2012-01-12 16:18:15 UTC

bugs in verified or on_qa moving off tracker

Comment 8 James Laska 2012-01-16 20:37:43 UTC

(In reply to comment #7)
> bugs in verified or on_qa moving off tracker

Verified and confirmed that ec2 image build+push succeeds when using a system template that specifies on-promise repos and packages.

aeolus-all-0.8.0-5.el6.src.rpm
aeolus-conductor-0.8.0-5.el6.src.rpm
aeolus-configure-2.5.0-3.el6.src.rpm
imagefactory-1.0.0rc2.1-1.el6.src.rpm
iwhd-1.2-3.el6.src.rpm
oz-0.7.9-7.el6.src.rpm
rubygem-aeolus-cli-0.3.0-3.el6.src.rpm
rubygem-aeolus-image-0.3.0-2.el6.src.rpm