Some time back the expected value of the openstack_repo_url
parameter was changed to make it more flexible.  The update was
done in such a way that the puppet-coe module would detect if the
old format was used and correct it on the fly, emitting a warning
in the process to let the user know what it did.  A composition
layer change was made to set the new format in
ebfc0a2, but only in
data/hiera_data/vendor/cisco_coi_common.yaml.  However, the value
is also explicitly set for HA scenarios in
data/hiera_data/enable_ha/true.yaml and that file was never
updated with the new format.  Thus, users of HA scenarios see
a warning emitted.  This patch updates the setting for HA
scenarios so the warning goes away.

Change-Id: I13bc10ea4ccbc064e8d76468b6bdae9df599f2d8
Closes-Bug: #1292152

Drop calls of naginator module classes to configure nagios by default as
the module design does not align with Puppet workflows as used with
stackforge modules, and as most users have preferred NMS solutions which
they use instead of the default example Nagios setup this created.

Change-Id: I51ba18b7c059a07e686ef0f2410d266ef08a37b8
Implements: blueprint nagios-ectomy
(cherry picked from commit 9e25883)

The region data mappings contain a couple of typos (single ':'
instead of two separating class names), and a few more need to be
added for classes that have added region params over time.  This
patch corrects errant mappings and adds missing ones.

Thanks to Ken Schroeder for the find and suggested fix.

Change-Id: I5fc5ae1b41ddc3fc5ddf5d7a481a5191a590509e
Closes-Bug: #1308366
(cherry picked from commit fbc939d)

When cobbler is enabled in the AIO scenario, there is a potential
duplicate resource conflict when defining hosts in the /etc/hosts
file.  Puppet-coi and puppet-coe both create host definitions
(one for the control node and one for the build node...which for
AIO are actually the same).  Install.sh creates a user.yaml when
called with the all_in_one scenario set and populates the control
node addresses into it...but doesn't populate cobbler_node_ip.
This can lead to a duplicate resource error because although the
two modules both use ensure_resource, the IP addresses aren't set
consistently in user.yaml for the AIO scenario.  This patch corrects
the problem by having install.sh set cobbler_node_ip to the same
IP address used for the controller.

Change-Id: Iafbe9b2603ab782d1ab8fd666312ca14bb6dd7f0
Closes-Bug: #1308257
(cherry picked from commit 9bba693)

This commit resolves -
* Issue where Apache is not listening on right
  bind address
* Issue where nova novpc should bind to a physical
  address instead of vip.

Fixes bug#1307536 & bug#1307537

Change-Id: I6f64e99ec7b62cff99bfec9666ca91655580a35c

fix the class parameter typo in true.yaml
bind address for horizon.

Change-Id: Icce56e43e6742b90851c148bc723a1b7005f7a54

I5fc5ae1b41 added several missing region data mappings.
However, one further mapping is also needed: we don't currently
have a mapping for the region setting in the ceilometer agent.
This patch adds a mapping for the
ceilometer::agent::auth::auth_region parameter.

Change-Id: I66e890737bbbd00a296e0bfa861ca8992b472c17
Closes-Bug: #1308366

master will probably bring icehouse specific changes which might break
in havana.

Change-Id: I4193c00e773f45561da9a9ce1dcf373b48648946

Cinder could only attach volumes to instances that were on the
same node. This is due to the iscsi listening address set to
0.0.0.0. Attaching the listener to a interface resolves the
problem and allows volume operations from/to any host.

Change-Id: Ia66c4f7f7278e2056d1ca3770ed2461259cca4e8
(cherry picked from commit 7b2e70a)