Advanced network support

This section covers the available features and how to configure them.

Load balancers

Blueprints can include load balancing to support easy scale up and down of their resulting resources.

CMP ships with load balancer support for AWS Elastic Load Balancing (ELB), F5 Big-IP, HAProxy, VMware NSX, and Microsoft Azure. Contact Support to request other technologies.

Create

To create a scalable blueprint with a load balancer, go to Catalog, add a blueprint, and add at least one server tier blueprint item and a load balancer blueprint item. The specific steps depend on the type of load balancer added.

AWS Elastic Load Balancer

CMP provisions “Network” Elastic Load Balancers.

No additional configuration is required for ELBs. EC2 instances in a server tier can be load-balanced and later scaled up by resource managers as needed.

1. Add a server tier blueprint item with an environment.
2. Add an ELB load balancer blueprint item, configure port settings, and select the tier to balance.

F5 Big-IP load balancer

This section describes how to use the F5 Big-IP integration that comes with CMP. To learn more about modifying load-balancing orchestration actions, contact Support.

F5 prerequisites

• Connect CMP to your F5 Big-IP appliance(s):
  • Go to Admin > Load Balancers.
  • Select Add and choose F5 Big-IP.
  • Enter the connection information for your F5 Big-IP.
  • Select Create.
• From the details page for the newly created F5 Big-IP, create an IP pool.

After completing the prerequisites, add an F5 load balancer to your blueprints:

1. Add a server tier blueprint item with an environment.
2. Add an F5 load balancer blueprint item, configure port settings, and select the tier to balance.
3. Select which F5 Big-IP appliance to use in the API calls to F5 Big-IP.

HAProxy load balancer

1. Add a server-tier blueprint item that provisions a single Linux server (Red Hat or CentOS) to run HAProxy.
2. Add a server tier blueprint item that provisions the servers you want load balanced.
3. Add an HAProxy load balancer blueprint item, configure port settings, and select the tier to balance.
4. Include the HAProxy tier (step 1), and choose whether CMP should use yum to install haproxy.

VMware NSX load balancer

If the server tier being balanced is deployed on a virtual network (NSX VXLAN), the NSX Edge Services Gateway that routes traffic to the VXLAN is automatically configured with a load balancer service. The only requirement is that the provider network for the edge appliance has an IP pool, so an IP can be auto-assigned to the load balancer service.

Scaling

Enable the Scale resource action, then select it from the details of any deployed resource to add or remove servers. If the target tier is marked as load balanced, servers are automatically added to or removed from the load balancer as needed.

Delete

To delete a load balancer, open your deployed resources and select the resource with the load balancer. On the resource’s Overview tab, select Delete. This deletes the resource and the load balancer.

Virtual networks

Blueprints can create virtual networks, providing server tiers with on-the-fly dedicated networks isolated from any other deployment of that blueprint.

CMP ships with virtual network support for VXLANs on vCenter instances configured with NSX. Virtual networks created from CMP are automatically associated with the vCenter resource handler.

Create

When creating a virtual network definition, assign a CIDR block and specify which server tier the network applies to. The network blueprint item must be placed before the target server tier in the blueprint deployment order.

Servers deployed as part of a targeted server tier are statically IP-configured, using IPs from a CMP-managed pool created at the same time as the network, based on the IP block definition.

The virtual network is named based on the resource’s deployment name and the server tiers it targets.

Routing virtual networks

When creating one or more virtual networks as part of a blueprint deployment, CMP automatically creates routing for the network, attaching the VXLANs to an NSX Edge Services Gateway (ESG) in one of two ways:

1. Attach to an existing ESG. If the environment where the server tier associated with the VXLAN has a single ESG option (check the VMware Parameters tab for the environment, NSX Edge parameter options), CMP attaches the VXLANs to internal interfaces on the ESG appliance. The ESG is expected to be configured with any desired routing to and from VXLANs outside of CMP.
2. Create a dedicated ESG. Alternatively, CMP can automatically create an ESG, defining a default static route and a basic SNAT rule to enable routing to and from the VXLANs. Creating ESGs depends on an NSX Edge Configuration object that specifies the datastore, provider network, edge CLI credentials, and other attributes used to create the ESG. See NSX for more on creating the NSX Edge Configuration object.

Scale up

When scaling up a resource tier that uses a virtual network, new servers in the tier are automatically assigned an IP from the CMP-managed pool and placed in the correct network. No additional setup is necessary.

Delete

When deleting a deployed resource that contains virtual networks, CMP sends the appropriate API request to the NSX Manager to delete the VXLANs. Deleting a virtual network requires it to be detached from any servers and routers; CMP reports any errors it encounters if the network cannot be deleted at that time. If you associate servers or routers with your VXLAN outside of CMP, refer to the vCenter instructions for deleting a VXLAN network.