I made 2 more attempts and installed the delorean release on bare metal:

1st install: AIO

2nd install: 2 node (controller + compute)

By the AIO install the dashboard worked fine.

By the 2nd install, I got "Bad Request (400)" by accessing the dashboard and it seems that the DocumentRoot in horizon vhost generated by puppet is not correct, so I changed it in:

/etc/httpd/conf.d/15-horizon_vhost.conf

  ## Vhost docroot

  #DocumentRoot "/var/www/"

  DocumentRoot "/usr/share/openstack-dashboard/static/"

Now the dashboard is working here too.

(if I re-run packstack the vhost will get overwritten again with the wrong docroot)

Another issue, after spawning a new cirros instance, it goes into error state and on the compute node I'm getting:

[root@compute1 ~]# tail -f /var/log/nova/nova-compute.log 

2015-04-26 11:17:16.027 13142 AUDIT nova.compute.resource_tracker [-] Total usable vcpus: 8, total allocated vcpus: 0

2015-04-26 11:17:16.027 13142 AUDIT nova.compute.resource_tracker [-] PCI stats: []

2015-04-26 11:17:16.063 13142 WARNING nova.compute.resource_tracker [-] No service record for host compute1

But the compute1 is enabled:

[root@csky06 ~(keystone_admin)]# nova service-list

+----+------------------+----------------------+----------+---------+-------+----------------------------+-----------------+

| Id | Binary           | Host                 | Zone     | Status  | State | Updated_at                 | Disabled Reason |

+----+------------------+----------------------+----------+---------+-------+----------------------------+-----------------+

| 1  | nova-consoleauth | controller | internal | enabled | up    | 2015-04-26T15:23:41.000000 | -               |

| 2  | nova-scheduler   | controller  | internal | enabled | up    | 2015-04-26T15:23:41.000000 | -               |

| 3  | nova-conductor   | controller | internal | enabled | up    | 2015-04-26T15:23:40.000000 | -               |

| 5  | nova-cert        | controller  | internal | enabled | up    | 2015-04-26T15:23:40.000000 | -               |

| 6  | nova-compute     | compute1 | nova     | enabled | up    | 2015-04-26T15:23:39.000000 | -               |

+----+------------------+----------------------+----------+---------+-------+----------------------------+-----------------+

And if I call nova hypervisor-list, the list is empty

[root@controller (keystone_admin)]# nova hypervisor-list

+----+---------------------+

| ID | Hypervisor hostname |

+----+---------------------+

+----+---------------------+

Not sure if that has something to do with the empty hypervisor list (?).

Thanks!