This guide details the usage of the Tuskar UI. It assumes that you have already installed Tuskar and the Tuskar UI. If you have not, please follow one of the Installation Guides first.
A narrated video is also available.
Start by logging into the Tuskar UI. The Tuskar UI is designed as a plugin to Horizon. The recommended installation will eliminate the standard 'Project' and 'Admin' dashboard, leaving only the Tuskar UI's 'Infrastructure' dashboard.
First you'll need to register some nodes. Click on the Nodes tab, then on the "Register Nodes" button in the upper-right corner.
You'll see a form that asks you to input your nodes' details. At a minimum you must specify the NIC mac address, the number of CPUs, the memory, and the local disk. You can enter information for multiple nodes by clicking on the "Add Node" link. Once done, click "Register Nodes".
If the nodes are registered successfully, you'll be back at the Nodes index page; only now there will be a count of Free Nodes. Before using these nodes for a cloud deployment, you'll want to wait a minute for nova to register them for use as a hypervisor.
The next step is to create flavors. Flavors are used so that your OpenStack Deployment knows which Nodes are appropriate for use.
Click on the Flavors tab. If no flavors exist, click on Flavor Suggestions. Here, you'll see suggested flavor specifications that match the nodes created in the previous step. To create them, simply click on the 'Create' button, and you'll be taken to a pre-filled flavor creation form. If you are satisfied with the form values, click on 'Create Flavor'. Make sure the "Deploy Kernel Image" and "Deploy Ramdisk Image" are set to the bm-deploy-kernel and bm-deploy-ramdisk images and the architecture is set to amd64.
Now if you select the 'Flavor' tab, you'll see that your flavor has been created.
Associate Flavors with Roles
Now that your nodes and flavors have been created, we can get started on creating the actual OpenStack Deployment. Click on that tab.
You'll see a deployment creation workflow, beginning with a list of pre-defined roles. However before specifying the node count for each role, you must specify the flavor that each role will use to match appropriate nodes. To do this for each role, click on 'Add a flavor'.
An 'Edit Deployment Role' form will pop up. As these are default roles, the only value that can be modified is Flavor. Select your desired flavor and click on 'Apply Changes'.
Specify Deployment Parameters and Options
Now it's time to specify the parameters and options for your Deployment. First, select the number of nodes you wish deployed in each role. Note that the Controller role must be exactly 1.
Afterwards, click on the Configuration tab. Here you can specify various configuration options for your deployment. You can also choose not to modify any of these values; Tuskar will use appropriate defaults. You'll need to set the NovaComputeLibvirtType flag to the correct option for your environment - probably kvm for Baremetal deployments and qemu for virtualized ones.
Once you are satisfied with your deployment choices, click on the 'Deploy' button. If everything goes as expected, you'll see a deployment progress bar, and here your watch begins.
As nodes are provisioned with the appropriate image for a role, this page will refresh with updated information. You can also click on the 'Log' tab to see a log of events that Heat is undertaking while creating the deployment stack.
For the installation via Devtest, the initialization code can be found in Overcloud deployment docs: http://docs.openstack.org/developer/tripleo-incubator/devtest_overcloud.html starting with step 13