Bug 1228165

It looks like the migration didn't work.  There's this:
   db/migrate/20150423134004_add_content_host_id_to_smart_proxy.rb 

Which is supposed to add the content_host to the smart proxy, and that didn't happen for some reason on your Satellite.

Post-upgrade I didn't encounter this issue, my Capsule had the Content Host associated.

However, deleting the Content Host associated with the Capsule, and later reregistering it will cause this problem, so its something we need to prevent from happening.

I've now done a bunch of upgrades with SNAP 7 through various bug testing and never hit this, so I don't think it's an upgrade issue, or if it was, its been corrected.

The situation could still happen if you unregister the content host that the capsule is attached to.  You can fix it in `foreman-rake console` with this:

User.current = User.anonymous_admin
 
SmartProxy.all.each do |proxy|
  content_host = ::Katello::System.where(:name => proxy.name).order("created_at DESC").first
 
  if content_host
    proxy.content_host_id = content_host.id
    proxy.save!
    puts "Capsule #{proxy} has Content Host #{content_host}"
  else
    puts "No Content Host Found for #{proxy}"
  end
end 


Maybe that could be turned into a rake task.


The permanent solution isn't clear to me, there's a few things we could do. Maybe we should delete the SmartProxy object if a user deletes the relevant ContentHost for a Capsule.

This issue is still reproducible on rhel66 with sat6.1.1 snap12.

here is the summary:
========================


While trying upgrade from sat6.0.8 to sat6.1.1 snap12, found a weird issue where content-view promotion ( Library -> dev) failed. Please note that CV publish was successful. I've updated the CV with 6.1.1 capsule and sat-tools repos. So I was trying to generate new version of CV with these repos.

Here is the error I found in production.log:

---
2015-07-13 03:25:17 [E] Could not find Content Host with exact name 'cloud-qe-20.idmqe.lab.eng.bos.redhat.com', verify the Capsule is registered with that name. (Katello::Errors::CapsuleContentMissingConsumer)
--

here 'cloud-qe-20.idmqe.lab.eng.bos.redhat.com' is the node which was configured as external capsule. I skipped the promotion task from dynflow and resume it again.

Later,  I tried capsule-upgrade and it was completed successfully (with workaround[1]). But when I initiated capsule sync, it got failed with same error:

--
[root@cloud-qe-17 ~]# hammer -u admin -p changeme capsule content synchronize --id=2
Could not synchronize capsule content:
  Could not find Content Host with exact name 'cloud-qe-20.idmqe.lab.eng.bos.redhat.com', verify the Capsule is registered with that name.
--

I unregister and re-register the capsule node again and re-initiated sync. this time sync successfully completed.

----
[root@cloud-qe-17 ~]# hammer -u admin -p changeme capsule content synchronize --id=2
[............................................................................................................................................] [100%]
-----

Created attachment 1051372 [details]
logs from production.log

Created attachment 1051398 [details]
CV promotion failed after sat6 upgrade

This issue is reproducible on rhel71 too.

steps performed:
===================

- I installed sat6.0.8 with external capsule and provisioned a content-host via external capsule.

- upgraded sat6.0.8 with sat6.1.1.  (upgrade was completed successfully)
- created custom repos for sat6.1_tools and capsule repo.
- added repos to existing CV and published it
- but when I promoted the new version to next 'dev' env. promotion failed with error:

--
2015-07-13 08:48:02 [E] Could not find Content Host with exact name 'cloud-qe-6.idmqe.lab.eng.bos.redhat.com', verify the Capsule is registered with that name. (Katello::Errors::CapsuleContentMissingConsumer)
/opt/rh/ruby193/root/usr/share/gems/gems/katello-2.2.0.57/app/lib/katello/capsule_content.rb:48:in `consumer'
/opt/rh/ruby193/root/usr/share/gems/gems/katello-2.2.0.57/app/lib/actions/katello/capsule_content/configure_capsule.rb:19:in `block in plan'
--

Created attachment 1051400 [details]
foreman-debug from rhel71

Sachin will open a separate docs BZ for the problem, will leave this one as-is, deleting a content host breaks synchronization.

Earlier I was trying satellite6.1 content synchronization (capsule repo and sat-tools repo) after upgrade. Which breaks the content promotion.

As suggested, this time I synced all repos(capsule and sat-tools) before upgrade. Upgrade on Satellite server and external capsule completed successfully.

But when I re-tried the capsule sync after upgrade, I got this:

[root@cloud-qe-12 ~]# hammer -u admin -p changeme capsule content synchronize --id=2
Could not synchronize capsule content:
  Could not find Content Host with exact name 'cloud-qe-15.idmqe.lab.eng.bos.redhat.com', verify the Capsule is registered with that name.

[root@cloud-qe-12 ~]#  hammer -u admin -p changeme capsule list
---|------------------------------------------|------------------------------------------------------
ID | NAME                                     | URL                                                  
---|------------------------------------------|------------------------------------------------------
1  | cloud-qe-12.idmqe.lab.eng.bos.redhat.com | https://cloud-qe-12.idmqe.lab.eng.bos.redhat.com:9090
2  | cloud-qe-15.idmqe.lab.eng.bos.redhat.com | https://cloud-qe-15.idmqe.lab.eng.bos.redhat.com:9090
---|------------------------------------------|------------------------------------------------------
[root@cloud-qe-12 ~]# 

Does this mean we need to re-register the capsule after upgrade ?

Created redmine issue http://projects.theforeman.org/issues/11123 from this bug

Moving to POST since upstream bug http://projects.theforeman.org/issues/11123 has been closed
-------------
Anonymous
Applied in changeset commit:katello|0bc31483f1110205dfa38ce1b0e16ce88dfd0839.

commit c0ca95be6ce42a9569010c4d0bc095afb732d89a
Author: Stephen Benjamin <stbenjam>
Date:   Wed Jul 15 10:41:14 2015 -0400

    fixes #11123 - actually assign content_host to capsule

Verified with sat6.1.1 GA snap14.

Now I can successfully sync contents to capsule after upgrade.

steps performed:

Upgraded sat server and capsule from  sat6.0.8 -> sat6.1.1 GA snap14
re-run the capsule-sync and it was completed without any error.

[root@cloud-qe-19 ~]# hammer -u admin -p changeme capsule content synchronize --id=2
[............................................................................................................................................] [100%]

A workaround for anyone encountering this until GA:

echo 'User.current = User.anonymous_admin; SmartProxy.all.each { |proxy| proxy.content_host.blank? ? proxy.update_attribute(:content_host, Katello::System.find_by_name(proxy.name)) : nil }' | foreman-rake console

This bug was fixed in Satellite 6.1.1 which was delivered on 12 August, 2015.