Enable Fasttrack

Defaults bifrost to leverage fast track. This will enable
bifrost's CI to identify failures related to code changes
around fast track... which is ultimately a little more
difficult to test in the OpenStack oriented CI jobs.

Change-Id: I643d4d3ee9f75cc7525ddf0551a03483906c433b

playbooks/roles/bifrost-ironic-install/README.md

@@ -160,6 +160,10 @@ inspector_log_dir: String value, default undefined. If set, it specifies a
 nginx_log_dir: String value, default /var/log/nginx. It specifies a log
                directory for nginx.
 
+fast_track: Boolean setting to enable ironic to leverage an already running
+            agent on the host being deployed such that deployment can begin
+            immediately as opposed to waiting for a complete system reboot.
+
 ### Hardware Inspection Support
 
 Bifrost also supports the installation of ironic-inspector in standalone

playbooks/roles/bifrost-ironic-install/defaults/main.yml

@@ -2,6 +2,11 @@
 # Cleaning turns on ironic conductor clean_nodes flag
 # which causes the nodes to be wiped after deletion.
 cleaning: false
+
+# Enable fast_track mode for ironic which allows skipping a power
+# cycle
+fast_track: true
+
 http_boot_folder: /httpboot
 ironic_tftp_master_path: /var/lib/ironic/master_images
 staging_drivers_include: false

playbooks/roles/bifrost-ironic-install/templates/ironic.conf.j2

@@ -62,6 +62,7 @@ uefi_pxe_config_template = $pybasedir/drivers/modules/ipxe_config.template
 http_url = http://{{ hostvars[inventory_hostname]['ansible_' + ans_network_interface]['ipv4']['address'] }}:{{ file_url_port }}/
 http_root = {{ http_boot_folder }}
 default_boot_option = local
+fast_track = {{ fast_track }}
 
 [conductor]
 clean_nodes = {{ cleaning | lower }}

releasenotes/notes/default-to-use-fasttrack-a526e1d86d2f669d.yaml

@@ -0,0 +1,8 @@
+---
+features:
+  - |
+    The default for bifrost is to enable ironic's fast_track mode using
+    ironic.conf's ``[deploy]fast_track`` option which enables ironic
+    to skip a power cycle sequence for deployments if the node power
+    is already on and the agent is running, which is how stand-alone
+    deployments tend to operate.