This patch provides a set of templates that enables tripleo-heat-templates to be used with a set of already deployed, installed, and running servers. In this method, Nova and Ironic are not used to deploy any servers. This approach is attractive for POC deployments where dedicated provisioning networks are not available, or other server install methods are dictated for various reasons. There are also assumptions that currently have to be made about the software installed on the already deployed servers. Effectively, they must match the standard TripleO overcloud-full image. Co-Authored-By: Steve Hardy <shardy@redhat.com> Change-Id: I4ab1531f69c73457653f1cca3fe30cc32a04c129
5.5 KiB
TripleO with Deployed Servers
The deployed-server set of templates can be used to deploy TripleO via tripleo-heat-templates to servers that are already installed with a base operating system.
When OS::TripleO::Server is mapped to the deployed-server.yaml template via the provided deployed-server-environment.yaml resource registry, Nova and Ironic are not used to create any server instances. Heat continues to create the SoftwareDeployment resources, and they are made available to the already deployed and running servers.
Template Usage
To use these templates pass the included environment file to the deployment command:
-e deployed-server/deployed-server-environment.yaml
Deployed Server configuration
It is currently assumed that the deployed servers being used have the required set of software and packages already installed on them. These exact requirements must match how such a server would look if it were deployed the standard way via Ironic using the TripleO overcloud-full image.
An easy way to help get this setup for development is to use an overcloud-full image from an already existing TripleO setup. Create the vm's for the already deployed server, and use the overcloud-full image as their disk.
Each server must have a fqdn set that resolves to an IP address on a routable network (e.g., the hostname should not resolve to 127.0.0.1). The hostname will be detected on each server via the hostnamectl --static command.
Each server also must have a route to the configured IP address on the undercloud where the OpenStack services are listening. This is the value for local_ip in the undercloud.conf.
It's recommended that each server have at least 2 nic's. One used for external management such as ssh, and one used for the OpenStack deployment itself. Since the overcloud deployment will reconfigure networking on the configured nic to be used by OpenStack, the external management nic is needed as a fallback so that all connectivity is not lost in case of a configuration error. Be sure to use correct nic config templates as needed, since the nodes will not receive dhcp from the undercloud neutron-dhcp-agent service.
For example, the net-config-static-bridge.yaml template could be used for controllers, and the net-config-static.yaml template could be used for computes by specifying:
- resource_registry:
-
OS::TripleO::Controller::Net::SoftwareConfig: /home/stack/deployed-server/tripleo-heat-templates/net-config-static-bridge.yaml OS::TripleO::Compute::Net::SoftwareConfig: /home/stack/deployed-server/tripleo-heat-templates/net-config-static.yaml
In a setup where the first nic on the servers is used for external management, set the nic's to be used for OpenStack to nic2:
- parameter_defaults:
-
NeutronPublicInterface: nic2 HypervisorNeutronPublicInterface: nic2
The above nic config templates also require a route to the ctlplane network to be defined. Define the needed parameters as necessary for your environment, for example:
- parameter_defaults:
-
ControlPlaneDefaultRoute: 192.168.122.130 ControlPlaneSubnetCidr: "24" EC2MetadataIp: "192.0.2.1"
In this example, 192.168.122.130 is the external management IP of an undercloud, thus it is the default route for the configured local_ip value of 192.0.2.1.
os-collect-config
os-collect-config on each deployed server must be manually configured to poll the Heat API for the available SoftwareDeployments. An example configuration for /etc/os-collect-config.conf looks like:
[DEFAULT] collectors=heat command=os-refresh-config
[heat] # you can get these values from stackrc on the undercloud user_id=<a user that can connect to heat> # note this must be the ID, not the username password=<a password> auth_url=<keystone url> project_id=<project_id> # note, this must be the ID, not project name stack_id=<stack_id> resource_name=<resource_name>
Note that the stack_id value is the id of the nested stack containing the resource (identified by resource_name) implemented by the deployed-server.yaml templates.
Once the configuration for os-collect-config has been defined, the service needs to be restarted. Once restarted, it will start polling Heat and applying the SoftwareDeployments.
A sample script at deployed-server/scripts/get-occ-config.sh is included that will automatically generate the os-collect-config configuration needed on each server, ssh to each server, copy the configuration, and restart the os-collect-config service.
Warning
The get-occ-config.sh script is not intended for production use, as it copies admin credentials to each of the deployed nodes.
The script can only be used once the stack id's of the nested deployed-server stacks have been created via Heat. This usually only takes a couple of minutes once the deployment command has been started. Once the following output is seen from the deployment command, the script should be ready to run:
[Controller]: CREATE_IN_PROGRESS state changed [NovaCompute]: CREATE_IN_PROGRESS state changed
The user running the script must be able to ssh as root to each server. Define the hostnames of the deployed servers you intend to use for each role type:
export controller_hosts="controller0 controller1 controller2"
export compute_hosts="compute0"
Then run the script on the undercloud with a stackrc file sourced, and the script will copy the needed os-collect-config.conf configuration to each server and restart the os-collect-config service.