xci: Updates to releng/bifrost to make it work on Jenkins

- get rid of BAREMETAL_DATA_FILE and use BIFROST_INVENTORY_SOURCE
so it works both for master and ocata for osa-bifrost.
- set BIFROST_INVENTORY_SOURCE file according to branch bifrost is
executed from for osa-bifrost.
- explicitly set what the SSH public key file is as ansible copies
the wrong public key to nodes if bifrost is executed on Jenkins
using sudo.
- set branches if they are not set so the scripts can be used manually
as well.
- rename jumphost to xcimaster so people do not mix it with actual
jumphost located in Pharos PODs.

Change-Id: Iff7631fa99816ad75316b62c5ac20714f67cd86a
Signed-off-by: Fatih Degirmenci <fatih.degirmenci@ericsson.com>

1 files changed, 53 insertions, 0 deletions

diff --git a/bifrost/playbooks/inventory/group_vars/baremetal b/bifrost/playbooks/inventory/group_vars/baremetal
new file mode 100644
index 00000000..008b04d1
--- /dev/null
+++ b/bifrost/playbooks/inventory/group_vars/baremetal
@@ -0,0 +1,53 @@
+---
+# The ironic API URL for bifrost operations.  Defaults to localhost.
+# ironic_url: "http://localhost:6385/"
+
+# The network interface that bifrost will be operating on.  Defaults
+# to virbr0 in roles, can be overridden here.
+# network_interface: "virbr0"
+
+# The path to the SSH key to be utilized for testing and burn-in
+# to configuration drives. When set, it should be set in both baremetal
+# and localhost groups, however this is only an override to the default.
+
+# workaround for opnfv ci until we can fix non-root use
+ssh_public_key_path: "/root/.ssh/id_rsa.pub"
+
+# Normally this user should be root, however if cirros is used,
+# a user may wish to define a specific user for testing VM
+# connectivity during a test sequence
+testing_user: root
+
+# The default port to download files via.  Required for IPA URL generation.
+# Presently the defaults are located in the roles, however if changed both
+# the localhost and baremetal group files must be updated.
+# file_url_port: 8080
+
+# IPA Image parameters.  If these are changed, they must be changed in
+# Both localhost and baremetal groups.  Presently the defaults
+# in each role should be sufficent for proper operation.
+# ipa_kernel: "{{http_boot_folder}}/coreos_production_pxe.vmlinuz"
+# ipa_ramdisk: "{{http_boot_folder}}/coreos_production_pxe_image-oem.cpio.gz"
+# ipa_kernel_url: "http://{{ hostvars[inventory_hostname]['ansible_' + network_interface]['ipv4']['address'] }}:{{file_url_port}}/coreos_production_pxe.vmlinuz"
+# ipa_ramdisk_url: "http://{{ hostvars[inventory_hostname]['ansible_' + network_interface]['ipv4']['address'] }}:{{file_url_port}}/coreos_production_pxe_image-oem.cpio.gz"
+
+# The http_boot_folder defines the root folder for the webserver.
+# If this setting is changed, it must be applied to both the baremetal
+# and localhost groups. Presently the role defaults are set to the value
+# below.
+# http_boot_folder: /httpboot
+
+# The settings for the name of the image to be deployed along with the
+# on disk location are below.  If changed, these settings must be applied
+# to both the baremetal and localhost groups.  If the file is already on
+# disk, then the image generation will not take place, otherwise an image
+# will be generated using diskimage-builder.
+# deploy_image_filename: "deployment_image.qcow2"
+# deploy_image: "{{http_boot_folder}}/{{deploy_image_filename}}"
+
+# Under normal circumstances, the os_ironic_node module does not wait for
+# the node to reach active state before continuing with the deployment
+# process.  This means we may have to timeout, to figure out a deployment
+# failed.  Change wait_for_node_deploy to true to cause bifrost to wait for
+# Ironic to show the instance in Active state.
+wait_for_node_deploy: false