Bug 1559703

It seems like this regression is introduced by this commit d7d435d59c0ae9eed36edfe0a0054f7f0fe0790c

index 687e166e7..2d8f12015 100644
--- a/app/models/katello/concerns/content_facet_host_extensions.rb
+++ b/app/models/katello/concerns/content_facet_host_extensions.rb
@@ -45,8 +45,12 @@ module Katello
         accepts_nested_attributes_for(
           :content_facet,
           self.nested_attributes_options[:content_facet].merge(
-            :reject_if => proc { |attributes| attributes['content_view_id'].blank? && attributes['lifecycle_environment_id'].blank? })
+            :reject_if => :content_facet_ignore_update?)
         )
+
+        def content_facet_ignore_update?(attributes)
+          self.content_facet.blank? && attributes.blank?

As in this case, if attributes is not blank(as some attributes are inherited from the host group) but content_view_id is blank, content_facet_ignore_update? will return false and then the validation on content_view_id is triggered which leads 
to the failure of creating a new host.

Another way to reproduce this.

On sat 6.3

foreman-console rake

irb(main):011:0> host = Host::Managed.new({"operatingsystem_id"=>3, "managed"=>"true", "name"=>"mod-ops01", "ip"=>"192.168.18.103", "hostgroup_id"=>6, "organization_id"=>1, "mac"=>"00:50:56:BF:2A:39", "architecture_id"=>1, "location_id"=>1, "domain_id"=>1})

irb(main):012:0> host.content_facet <===== On Sat 6.2, this returns nil

=> #<Katello::Host::ContentFacet id: nil, host_id: nil, uuid: nil, content_view_id: nil, lifecycle_environment_id: 5, kickstart_repository_id: nil, content_source_id: 1, installable_security_errata_count: 0, installable_enhancement_errata_count: 0, installable_bugfix_errata_count: 0, applicable_rpm_count: 0, upgradable_rpm_count: 0>
irb(main):005:0> host = Host::Managed.new({"operatingsystem_id"=>3, "managed"=>"true", "name"=>"mod-ops01", "ip"=>"192.168.18.103", "hostgroup_id"=>6, "organization_id"=>1, "mac"=>"00:50:56:BF:2A:39", "architecture_id"=>1, "location_id"=>1, "domain_id"=>1})

irb(main):018:0> host.validate
=> false
irb(main):019:0> host.errors.first
=> [:"content_facet.content_view", "can't be blank"]

(In reply to matt jia from comment #2)
> Another way to reproduce this.
> 
> On sat 6.3
> 
> foreman-console rake

This should be foreman-rake console, :-)

(In reply to matt jia from comment #1)
> It seems like this regression is introduced by this commit
> d7d435d59c0ae9eed36edfe0a0054f7f0fe0790c
> 
> index 687e166e7..2d8f12015 100644
> --- a/app/models/katello/concerns/content_facet_host_extensions.rb
> +++ b/app/models/katello/concerns/content_facet_host_extensions.rb
> @@ -45,8 +45,12 @@ module Katello
>          accepts_nested_attributes_for(
>            :content_facet,
>            self.nested_attributes_options[:content_facet].merge(
> -            :reject_if => proc { |attributes|
> attributes['content_view_id'].blank? &&
> attributes['lifecycle_environment_id'].blank? })
> +            :reject_if => :content_facet_ignore_update?)
>          )
> +
> +        def content_facet_ignore_update?(attributes)
> +          self.content_facet.blank? && attributes.blank?
> 
> As in this case, if attributes is not blank(as some attributes are inherited
> from the host group) but content_view_id is blank,
> content_facet_ignore_update? will return false and then the validation on
> content_view_id is triggered which leads 
> to the failure of creating a new host.

I may get wrong here as I'm quite sure what self.nested_attributes_options does here.

Hi Matt,

Do you have an internal reproducer?   If so, would you be interested in trying a quick patch?

I suspect that this issue may be solved by the following upstream issue http://projects.theforeman.org/issues/21670.

(In reply to Brad Buckingham from comment #6)
> Hi Matt,
> 
> Do you have an internal reproducer?   If so, would you be interested in
> trying a quick patch?
> 
> I suspect that this issue may be solved by the following upstream issue
> http://projects.theforeman.org/issues/21670.

Hi Brad,

Yeah, it should be fixed by the patch in this upstream issue.

(In reply to matt jia from comment #7)
> (In reply to Brad Buckingham from comment #6)
> > Hi Matt,
> > 
> > Do you have an internal reproducer?   If so, would you be interested in
> > trying a quick patch?
> > 
> > I suspect that this issue may be solved by the following upstream issue
> > http://projects.theforeman.org/issues/21670.
> 
> Hi Brad,
> 
> Yeah, it should be fixed by the patch in this upstream issue.

It turns out this patch doesn't solve this issue. I will update a patch then.

Created attachment 1412963 [details]
propose a patch

(In reply to matt jia from comment #2)
> Another way to reproduce this.
> 
> On sat 6.3
> 
> foreman-console rake
> 
> irb(main):011:0> host = Host::Managed.new({"operatingsystem_id"=>3,
> "managed"=>"true", "name"=>"mod-ops01", "ip"=>"192.168.18.103",
> "hostgroup_id"=>6, "organization_id"=>1, "mac"=>"00:50:56:BF:2A:39",
> "architecture_id"=>1, "location_id"=>1, "domain_id"=>1})
> 
> irb(main):012:0> host.content_facet <===== On Sat 6.2, this returns nil
> 
> => #<Katello::Host::ContentFacet id: nil, host_id: nil, uuid: nil,
> content_view_id: nil, lifecycle_environment_id: 5, kickstart_repository_id:
> nil, content_source_id: 1, installable_security_errata_count: 0,
> installable_enhancement_errata_count: 0, installable_bugfix_errata_count: 0,
> applicable_rpm_count: 0, upgradable_rpm_count: 0>
> irb(main):005:0> host = Host::Managed.new({"operatingsystem_id"=>3,
> "managed"=>"true", "name"=>"mod-ops01", "ip"=>"192.168.18.103",
> "hostgroup_id"=>6, "organization_id"=>1, "mac"=>"00:50:56:BF:2A:39",
> "architecture_id"=>1, "location_id"=>1, "domain_id"=>1})
> 
> irb(main):018:0> host.validate
> => false
> irb(main):019:0> host.errors.first
> => [:"content_facet.content_view", "can't be blank"]

To be clear, the host group used here does not associate with any content view.

Matt,

Thanks for the feedback.  I ran a similar test on my 6.3.1 environment and I had different results.  On my environment, applying the patch from redmine 21670 solve the issue.

Console output from my test:

BEFORE PATCH

[root@sat-6-3-qa-rhel7 ~]# foreman-rake console
Loading production environment (Rails 4.2.6)
Failed to load console gems, starting anyway
irb(main):001:0> host=Host::Managed.new({"operatingsystem_id"=>2,"managed"=>"false","name"=>"testhost", "ip"=>"192.168.18.103","hostgroup_id"=>1,"organization_id"=>1, "mac"=>"00:50:56:BF:2A:39", "architecture_id"=>1,"location_id"=>1, "domain_id"=>1})
=> #<Host::Managed id: nil, name: "testhost", last_compile: nil, last_report: nil, updated_at: nil, created_at: nil, root_pass: nil, architecture_id: 1, operatingsystem_id: 2, environment_id: nil, ptable_id: nil, medium_id: nil, build: false, comment: nil, disk: nil, installed_at: nil, model_id: nil, hostgroup_id: 1, owner_id: nil, owner_type: nil, enabled: true, puppet_ca_proxy_id: nil, managed: false, use_image: nil, image_file: nil, uuid: nil, compute_resource_id: nil, puppet_proxy_id: nil, certname: nil, image_id: nil, organization_id: 1, location_id: 1, type: "Host::Managed", otp: nil, realm_id: nil, compute_profile_id: nil, provision_method: nil, grub_pass: "", discovery_rule_id: nil, content_view_id: nil, lifecycle_environment_id: nil, global_status: 0, lookup_value_matcher: nil, pxe_loader: nil, openscap_proxy_id: nil>
irb(main):002:0> host.validate
=> false
irb(main):003:0> host.errors.first
=> [:"content_facet.content_view", "can't be blank"]
irb(main):004:0> 
irb(main):005:0* 
irb(main):006:0* exit


AFTER PATCH

[root@sat-6-3-qa-rhel7 ~]# foreman-rake console
Loading production environment (Rails 4.2.6)
Failed to load console gems, starting anyway
irb(main):001:0> host=Host::Managed.new({"operatingsystem_id"=>2,"managed"=>"false","name"=>"testhost", "ip"=>"192.168.18.103","hostgroup_id"=>1,"organization_id"=>1, "mac"=>"00:50:56:BF:2A:39", "architecture_id"=>1,"location_id"=>1, "domain_id"=>1})
=> #<Host::Managed id: nil, name: "testhost", last_compile: nil, last_report: nil, updated_at: nil, created_at: nil, root_pass: nil, architecture_id: 1, operatingsystem_id: 2, environment_id: nil, ptable_id: nil, medium_id: nil, build: false, comment: nil, disk: nil, installed_at: nil, model_id: nil, hostgroup_id: 1, owner_id: nil, owner_type: nil, enabled: true, puppet_ca_proxy_id: nil, managed: false, use_image: nil, image_file: nil, uuid: nil, compute_resource_id: nil, puppet_proxy_id: nil, certname: nil, image_id: nil, organization_id: 1, location_id: 1, type: "Host::Managed", otp: nil, realm_id: nil, compute_profile_id: nil, provision_method: nil, grub_pass: "", discovery_rule_id: nil, content_view_id: nil, lifecycle_environment_id: nil, global_status: 0, lookup_value_matcher: nil, pxe_loader: nil, openscap_proxy_id: nil>
irb(main):002:0> host.validate
=> true
irb(main):003:0> host.errors.first
=> nil
irb(main):004:0> 


Note: on my test, I basically created a hostgroup with no details (e.g. no CV...etc) and then created the host using that group.

I tested Matt's second patch on a 6.3.1 production server and it worked fine.

We will produce a hotfix set of RPMs which should be available soon to resolve this issue.

(In reply to Brad Buckingham from comment #12)
> 
> [root@sat-6-3-qa-rhel7 ~]# foreman-rake console
> Loading production environment (Rails 4.2.6)
> Failed to load console gems, starting anyway
> irb(main):001:0>
> host=Host::Managed.new({"operatingsystem_id"=>2,"managed"=>"false",
> "name"=>"testhost",
> "ip"=>"192.168.18.103","hostgroup_id"=>1,"organization_id"=>1,
> "mac"=>"00:50:56:BF:2A:39", "architecture_id"=>1,"location_id"=>1,
> "domain_id"=>1})
> => #<Host::Managed id: nil, name: "testhost", last_compile: nil,
> last_report: nil, updated_at: nil, created_at: nil, root_pass: nil,
> architecture_id: 1, operatingsystem_id: 2, environment_id: nil, ptable_id:
> nil, medium_id: nil, build: false, comment: nil, disk: nil, installed_at:
> nil, model_id: nil, hostgroup_id: 1, owner_id: nil, owner_type: nil,
> enabled: true, puppet_ca_proxy_id: nil, managed: false, use_image: nil,
> image_file: nil, uuid: nil, compute_resource_id: nil, puppet_proxy_id: nil,
> certname: nil, image_id: nil, organization_id: 1, location_id: 1, type:
> "Host::Managed", otp: nil, realm_id: nil, compute_profile_id: nil,
> provision_method: nil, grub_pass: "", discovery_rule_id: nil,
> content_view_id: nil, lifecycle_environment_id: nil, global_status: 0,
> lookup_value_matcher: nil, pxe_loader: nil, openscap_proxy_id: nil>
> irb(main):002:0> host.validate
> => true
> irb(main):003:0> host.errors.first
> => nil
> irb(main):004:0> 
> 
> 
> Note: on my test, I basically created a hostgroup with no details (e.g. no
> CV...etc) and then created the host using that group.

Hi Brad,

Can you please set a life cycle environment on that group and try again? I forgot to mention that the group I used for reproducing has associated with
a life cycle environment.

Created redmine issue http://projects.theforeman.org/issues/23457 from this bug

Moving this bug to POST for triage into Satellite 6 since the upstream issue http://projects.theforeman.org/issues/23457 has been resolved.

Hi Mike/Team,

One of my customer is facing this issue after upgrading satellite to 6.3.2. Can i provide the same hotfix. Case no #02127577.

Regards,

** Hotfix RPMs for 6.3.2 **

The attached RPMs only work on Satellite 6.3.2. Instructions for hotfix:

1) download newest RPMs from BZ attachment:

tfm-rubygem-katello-3.4.5.74-2.HOTFIXRHBZ1559703.el7sat.noarch.rpm
tfm-rubygem-katello_ostree-3.4.5.74-2.HOTFIXRHBZ1559703.el7sat.noarch.rpm

2) rpm -Uvh tfm-rubygem-katello-3.4.5.74-2.HOTFIXRHBZ1559703.el7sat.noarch.rpm
tfm-rubygem-katello_ostree-3.4.5.74-2.HOTFIXRHBZ1559703.el7sat.noarch.rpm

NOTE: Remove ostree RPM from command if not already installed. 

3) katello-service restart

4) resume operations and verify hotfix is functional in your environment

Created attachment 1455088 [details]
tfm-rubygem-katello-3.4.5.74-2.HOTFIXRHBZ1559703.el7sat.noarch.rpm

Created attachment 1455090 [details]
tfm-rubygem-katello_ostree-3.4.5.74-2.HOTFIXRHBZ1559703.el7sat.noarch.rpm

Hi Mike,

My customer has applied the hotfix for 6.3.2 and it works like a charm.

Thanks,

Andrew,

This failed to cherry-pick downstream, can you take a look?

Verified!

@ Satellite 6.3.3 snap 2


Steps and observation:
---------------------------

irb(main):001:0> host5 = Host::Managed.new({"operatingsystem_id"=>2, "managed"=>"true", "name"=>"jyejare-10", "ip"=>"192.168.100.157", "hostgroup_id"=>1, "mac":"00:50:56:bf:2a:30", "architecture_id"=>1, "location_id"=>2, "domain_id"=>1, "organization_id"=>1})
=> #<Host::Managed id: nil, name: "jyejare-10", last_compile: nil, last_report: nil, updated_at: nil, created_at: nil, root_pass: nil, architecture_id: 1, operatingsystem_id: 2, environment_id: 1, ptable_id: nil, medium_id: nil, build: false, comment: nil, disk: nil, installed_at: nil, model_id: nil, hostgroup_id: 1, owner_id: nil, owner_type: nil, enabled: true, puppet_ca_proxy_id: 1, managed: true, use_image: nil, image_file: nil, uuid: nil, compute_resource_id: nil, puppet_proxy_id: 1, certname: nil, image_id: nil, organization_id: 1, location_id: 2, type: "Host::Managed", otp: nil, realm_id: nil, compute_profile_id: nil, provision_method: nil, grub_pass: "", discovery_rule_id: nil, content_view_id: nil, lifecycle_environment_id: nil, global_status: 0, lookup_value_matcher: nil, pxe_loader: nil, openscap_proxy_id: nil>
irb(main):002:0> host5.validate
=> true
irb(main):003:0> host5.errors.first
=> nil
irb(main):004:0> host5.hostgroup
=> #<Hostgroup id: 1, name: "HG1", created_at: "2018-07-31 16:13:36", updated_at: "2018-07-31 16:13:36", environment_id: 1, operatingsystem_id: 2, architecture_id: 1, medium_id: nil, ptable_id: 92, root_pass: "", puppet_ca_proxy_id: 1, use_image: nil, image_file: nil, ancestry: nil, vm_defaults: nil, subnet_id: 1, domain_id: 1, puppet_proxy_id: 1, title: "HG1", realm_id: nil, compute_profile_id: nil, content_source_id: 1, grub_pass: "", content_view_id: nil, lifecycle_environment_id: 10, lookup_value_matcher: "hostgroup=HG1", kickstart_repository_id: nil, subnet6_id: nil, pxe_loader: "PXELinux BIOS", openscap_proxy_id: nil>
irb(main):005:0> hg = host5.hostgroup
=> #<Hostgroup id: 1, name: "HG1", created_at: "2018-07-31 16:13:36", updated_at: "2018-07-31 16:13:36", environment_id: 1, operatingsystem_id: 2, architecture_id: 1, medium_id: nil, ptable_id: 92, root_pass: "", puppet_ca_proxy_id: 1, use_image: nil, image_file: nil, ancestry: nil, vm_defaults: nil, subnet_id: 1, domain_id: 1, puppet_proxy_id: 1, title: "HG1", realm_id: nil, compute_profile_id: nil, content_source_id: 1, grub_pass: "", content_view_id: nil, lifecycle_environment_id: 10, lookup_value_matcher: "hostgroup=HG1", kickstart_repository_id: nil, subnet6_id: nil, pxe_loader: "PXELinux BIOS", openscap_proxy_id: nil>
irb(main):006:0> hg.content_view_id
=> nil


Hence, its proved, The host is being created without CV.


The same has been verified with manage = false and its working fine.

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-2018:2551