Platform Tagging

This feature is part of TECHNICAL PREVIEW.

You can define platform tags and attach them to credentials, networks, and templates to bundle together different configurations.

Data Locality and Topologies

The OpenStack documentation says this about data locality: "It is extremely important for data processing to do locally (on the same rack, OpenStack compute node or even VM) as much work as possible. Hadoop supports data-locality feature and can schedule jobs to task tracker nodes that are local for input stream. In this case task tracker could communicate directly with local data node."

OpenStack Topology Mapping

You can create a topology mapping, which associates hypervisors with racks, and then attach it to the platform definition. You can set the mapping in the Cloudbreak UI or in CLI by defining it line by line or uploading the mapping in a file.

The mapping file should have the following format:

hypervisor1 /rack1
hypervisor2 /rack2
hypervisor3 /rack2

Based on this mapping, the Cloudbreak application ensures that the rack information of the started VMs will be passed to Hadoop services via Ambari.

Platform Configuration Through Cloudbreak UI

Manage Platform Configuration

You can log in to the Cloudbreak application at https://PUBLIC_IP. You can find the provider-specific documentation here:

To create a new platform configuration:

  1. Go to the manage platforms panel.
  2. Select OpenStack from cloud provider tabs.
  3. Fill out the new configuration Name.
  4. Complete the Topology Mapping section based on the above examples. Add mapping line by line or upload a mapping file.

Full size here.

NOTE: Platform name is the only required parameter.

Explanation of the parameters:

Required parameters:

Optional parameters:

Enable Platform Configuration

You can apply your Cloudbreak platform configuration in the following panels:

To create a new network with a previously created configuration, follow these steps in the create network panel:

  1. From cloud provider tabs, select OpenStack.
  2. Enter the new network Name.
  3. Enter your network Subnet (CIDR).
  4. Enter your Public Network ID.
  5. Uder Select Platform, select your previously created platform.

Full size here.

IMOPRTANT: If you assign a platform to a selected credential, then only networks and templates associated with that credential can be selected during cluster creation.

Platform Configuration via CLI

Manage Platform Configuration

Start the shell with cbd util cloudbreak-shell on a console. This will launch the Cloudbreak shell inside a Docker container and you can start using it. For more information, see Cloudbreak Shell.

Here are two examples: For AWS:

# Creating a Platform
platform create --AWS --name platform-name --description 'description of the platform'

For OPENSTACK:

# Creating OpenStack Platform with topology mapping
platform create --OPENSTACK --name platform-name --description 'openstack platform' --file file_path
platform create --OPENSTACK --name platform-name --description 'openstack platform' --url url_to_file

Enable Platform Configuration

You can use the following Cloudbreak shell commands to set a platform for a resource:

Here is an example shell command to create a new network with a connected platform:

network create --AWS --name aws-network --subnet 10.10.10.0/24 --description 'example network' --platformId 26

IMOPRTANT: If you assign a platform to a selected credential, then only networks and templates associated with that credential can be selected during cluster creation.

Edit on GitHub