Bug 1144194

Summary:	Repository distributions whose contents change cause distribution to be published ever again
Product:	Red Hat Satellite	Reporter:	Justin Sherrill <jsherril>
Component:	Content Management	Assignee:	Katello Bug Bin <katello-bugs>
Status:	CLOSED ERRATA	QA Contact:	Corey Welton <cwelton>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	6.0.4	CC:	abraverm, achan, bbuckingham, bthurber, cduryee, cwelton, jmontleo, mmccune, omaciel, sthirugn
Target Milestone:	Unspecified	Keywords:	Triaged
Target Release:	Unused
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:	When changing the repository distributions of a kickstart tree from one version to another, the repo sync failed during the pulp distributor publish step. Changes in pulp have fixed the issue and the resync should successfully execute.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2014-11-13 22:29:19 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:	1150714
Bug Blocks:

Description Justin Sherrill 2014-09-19 00:49:34 UTC

Description of problem:

In Satellite 6 if a user syncs the 6Server, 5Server or 7Server kickstart tree and for the synced release a new Kickstart tree is released, the contents of the old kickstart tree will change and will affect all repos using the same distribution.

For example:

If you sync the RHEL 5.10 and 5Server kickstart trees.  Both repos will use a  

ks-Red Hat Enterprise Linux Server-5.10-x86_64

When 5.11 is released 5Server now points to 5.11 and so when satellite syncs it, pulp updates the existing ks-Red Hat Enterprise Linux Server-5.10-x86_64 kickstart tree with 5.11 content.  As a result you have a 5.10 kickstart tree serving 5.11. Both repos (even the 5.10 repo) have a 5.11 kickstart tree.


How reproducible:
Always

Steps to Reproduce:
1.  Create 2 repos pointing to a kickstart tree of release X.y
2.  Sync both repos
3.  Change the distribution in one of the repos to X.y+1 and sync it

Actual results:
since the distribution is shared between repos it is updated for both repos

Expected results:
the 5Server repo's distribution is independent of 5.10's and 


Additional info:

This causes lots of issues on the 'new host' page when trying to provision a host.  selecting 5.10 causes nothing to show up as a kickstart url

Comment 3 Justin Sherrill 2014-09-23 02:47:13 UTC

Note that the original problem is no longer reproducable.  I believe the state of the kickstart trees on the CDN was in such a poor state as to cause the issue as described.

When i try to reproduce the issue as it should happen with valid kickstart trees, i get:

Sep 19 15:33:32 abed pulp: celery.worker.job:ERROR: Exception: Error publishing repository Default_Organization-el5_product-5Server.  More
than one distribution found.

so it seems that pulp is not actually disassociating the previous distribution from the repository when the new one is found.  Checking pulp, sure enough there are 2 distributions associated.

The results of this are that if a user is using a XServer or XClient kickstart tree and a new version of rhel is released, that kickstart tree will sync just fine, but will not publish properly ever again (without user intervention).

The way to reproduce this is:

1.  Create a repo with a feed pointing to a 6.3 kickstart tree
2.  Sync the repo
3.  Change the feed url to a 6.4 kickstart tree
4.  Sync the repo

The katello repo sync will fail during the pulp distributor publish step.

Comment 4 Chris Duryee 2014-10-08 14:59:26 UTC

taking this bz to investigate. I'll make a pulp bz w/ a depends-on relationship after I dig into this a bit

Comment 5 Chris Duryee 2014-10-08 19:26:09 UTC

unassigning bz from myself and taking upstream bz

Comment 6 Jason Montleon 2014-10-22 14:21:43 UTC

I hit this today. I was able to install these on the system:
pulp-admin-client-2.4.1-0.7.beta.el6sat.noarch
pulp-rpm-admin-extensions-2.4.1-0.7.beta.el6sat.noarch

The changed the following lines in /etc/pulp/admin/admin.conf:
host = <hostname>
ca_path = /root/ssl-build/katello-ca.crt

pulp-admin login -u <username> -p <password>

pulp-admin rpm repo content distribution --repo-id <id>

pulp-admin rpm repo remove distribution --repo-id <id> --in="id=id-from-last"

Went back to the UI and tried to re-sync and it appeared to work.

Comment 7 Jason Montleon 2014-11-03 14:59:47 UTC

I ran into another issue that I am not certain is a part of this or something separate. If separate I can file a different bug.

What resulted from having a RedHat 6.5 Operating system configured to use 6Server installation media and then hitting this bug and working around it was an Operating System booting from a RHEL 6.5 kernel and initrd while using 6.6 installation media.

This resulted in getting a kernel panic and error when trying to install:
VFS: Cannot open root device "(null)" or unknown-block(8,5)

Setting up the newly created RedHat 6.6 Operating System to use the installation media and appropriate templates worked correctly. But it seems we can cause possibly unexpected breakage when modifying 6Server content. So I'm not sure if it makes sense to force using only, i.e. 6.5, 6.6, 7.0, content for kickstarting, since it shouldn't change in ways that will break, or if the change can somehow be accounted for and settings updated to reflect the change (and if it already exists if it didn't work because of this bug).

Comment 8 Mike McCune 2014-11-04 17:32:24 UTC

moving to MODIFIED since Pulp bug is fixed in 2.4.2 and we are using 2.4.3

Comment 9 Mike McCune 2014-11-06 16:38:45 UTC

*** Bug 1158973 has been marked as a duplicate of this bug. ***

Comment 11 Corey Welton 2014-11-07 22:35:41 UTC

Verified in Satellite-6.0.4-RHEL-7-20141107.

Comment 13 errata-xmlrpc 2014-11-13 22:29:19 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://access.redhat.com/errata/RHBA-2014:1857

Comment 14 Mike McCune 2015-06-02 17:37:20 UTC

*** Bug 1221151 has been marked as a duplicate of this bug. ***