[Rdo-list] Can't Deploy Foreman with openstack-foreman-installer for Bare Metal Provisioning (undefined method `' for nil:NilClass)
ramon at linux-labs.net
Thu May 8 15:23:07 UTC 2014
On 8 May 2014, at 02:25, Gilles Dubreuil <gilles at redhat.com> wrote:
> ----- Original Message -----
>> From: "Ramon Acedo" <ramon at linux-labs.net>
>> To: rdo-list at redhat.com
>> Sent: Wednesday, 30 April, 2014 2:14:59 AM
>> Subject: [Rdo-list] Can't Deploy Foreman with openstack-foreman-installer for Bare Metal Provisioning (undefined
>> method `' for nil:NilClass)
>> Hi all,
>> I have been trying to test the OpenStack Foreman Installer with different
>> combinations of Foreman versions and of the installer itself (and even
>> different versions of Puppet) with no success so far.
>> I know that Packstack alone works but I want to go all the way with multiple
>> hosts and bare metal provisioning to eventually use it for large deployments
>> and scale out Nova Compute and other services seamlessly.
>> The error I get when running the foreman_server.sh script is always:
>> rake aborted!
>> undefined method `' for nil:NilClass
>> Tasks: TOP => db:seed
>> (See full trace by running task with --trace)
> The above usually indicates there is something wrong with at least one puppet class.
> Do you have openstack-puppet-modules installed?
> Some of the devil's details:
> foreman_server.sh triggers foreman's rake to seed its database.
> The nil:NiClass means something is missing and usually when it happens, to be confirmed with rake's trace/logs it's because at least one puppet class is wrong (not validated).
> The above is happening because seeding script also parses puppet classes' parameters in order to inject them into Foreman.
After modifying the script to use "rake —trace” this is what I got:
+ sudo -u foreman scl enable ruby193 'cd /usr/share/foreman; rake --trace db:seed RAILS_ENV=production FOREMAN_PROVISIONING=true'
** Invoke db:seed (first_time)
** Execute db:seed
** Invoke db:abort_if_pending_migrations (first_time)
** Invoke environment (first_time)
** Execute environment
** Invoke db:load_config (first_time)
** Execute db:load_config
** Execute db:abort_if_pending_migrations
undefined method `' for nil:NilClass
I’m still trying to further debug it.
>> After that, if Foreman starts, there’s nothing in the "Host groups" section
>> which is supposed to be prepopulated by the foreman_server.sh script (as
>> described in http://red.ht/1jdJ03q).
>> The process I follow is very simple:
>> 1. Install a clean RHEL 6.5 or CentOS 6.5
>> 2. Enable EPEL
>> 3. Enable the rdo-release repo:
>> a. rdo-release-havana-7: Foreman 1.3 and openstack-foreman-installer 1.0.6
>> b. rdo-release-havana-8: Foreman 1.5 and openstack-foreman-installer 1.0.6
>> c. rdo-release-icehouse-3: Foreman 1.5 and openstack-foreman-installer 2.0
>> (as a note here, the SCL repo needs to be enabled before the next step
>> 4. Install openstack-foreman-installer
>> 5. Create and export the needed variables:
>> export PROVISIONING_INTERFACE=eth0
>> export FOREMAN_GATEWAY=192.168.5.100
>> export FOREMAN_PROVISIONING=true
>> 6. Run the script foreman_server.sh from
>> For 3a and 3b I also tried with an older version of Puppet (3.2) with the
>> same result.
>> These are the full outputs:
>> 3a: http://fpaste.org/97739/ (Havana and Foreman 1.3)
>> 3b: http://fpaste.org/97760/ (Havana and Foreman 1.3 with Puppet 3.2)
>> 3c: http://fpaste.org/97838/ (Icehouse and Foreman 1.5)
>> I’m sure somebody in the list has tried to deploy and configure Foreman for
>> bare metal installations (DHCP+PXE) and the documentation and the
>> foreman_server.sh script suggest it should be possible in a fairly easy way.
>> I filled a bug as it might well be one, pending confirmation:
>> Any help is really appreciated!
>> Many thanks.
>> Rdo-list mailing list
>> Rdo-list at redhat.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the dev