OpenStack Discussions

OpenStack on FlexPod CVD Deployment Guide with NetApp, Cisco, and Red Hat


Hi Folks,

I’m happy to share with you the publishing of the second document of the FlexPod Datacenter Cisco Validated Design (CVD) with Red Hat Enterprise Linux OpenStack Platform 6 project, this time the Deployment Guide! Paired with it's partner Design Guide (released 8/3), this CVD represents an enterprise-grade open hybrid cloud foundation that lets you deploy OpenStack in an accelerated fashion with less risk.

Document links

Deployment Guide (NEW) -- is a step-by-step, detailed implementation guide giving full steps on implementing the solution. Also included is a subset of the Design Guide in order to provide necessary context for the Deployment Guide.

Design Guide -- typically includes design elements integral to the overall solution as well as why technology chosen in the architectural design represent the best choice for a deployment.

About these documents

This collaborative, technical engineering effort represents a fully supported, enterprise-class converged infrastructure deployment on FlexPod built with NetApp® FAS and E-Series storage, Cisco® UCS servers and Cisco Nexus switches, and Red Hat's Enterprise Linux OpenStack Platform.

In the deployment guide, you can find the following and more:

  1. Context (From the Design Guide) and Technology Overview as to why FlexPod represents the best enterprise-class converged infrastructure platform for OpenStack deployments.

  2. Detailed implementation instructions on how to configure a fresh out-of-the-box FAS8040 two-node pair redundantly and prepare it for usage for both:

    • Cinder volumes and Glance images utilizing NetApp technology exposed to it through the NetApp Unified Driver and the NetApp Clustered Data ONTAP Operating System
    • Stateless computing via iSCSI boot LUNs to eventual Cisco UCS Service Profiles
  3. Detailed implementation instructions on how to bring up and configure an E5560 dual-controller storage system and prepare it as the backbone for a highly available and resilient OpenStack Object Storage deployment.

  4. Deploy Cisco Unified Computing System (UCS) via Cisco UCS Manager (UCSM) to provide a robust and scalable platform for server infrastructure and compute for OpenStack:

    • Guidance on VLANs, Service Profile templates, vNIC placement and failover, and Uplink port channels to the Cisco Nexus 9000 switch pair.
    • Implementation instructions for common pools across the server infrastructure (MAC address, IQN, Server, Boot Policy, Firmware, etc)
  5. Step-by-step instructions on deploying the Red Hat Enterprise Linux OpenStack Platform installer, a lifecycle management and orchestration utility used to:

    • discover via PxE physical server blades ready to be used for OpenStack
    • provide a graphical user interface used to configure and launch an OpenStack Deployment
    • deploy Red Hat Enteprise Linux 7.1 on bare-metal UCS server blades
    • install all necessary OpenStack related packages via Puppet in an automated, HA deployment
    • take user input via the graphical user interface to take advantage of a NetApp FAS8040 and be ready for Cinder volumes immediately after the OpenStack deployment is finished, with the NetApp Unified Driver configured and active
  6. Post Deployment Instructions after the OpenStack deployment is finished:

    • Implement a Storage Service Catalog via Cinder to expose NetApp technology and value to the user directly to meet a diverse base of application and tenant needs.
    • Implement the NetApp Copy Offload Tool, an efficient way of copying Glance images to Cinder volumes through the storage system itself, rather than through the network as it's first copy.
    • Deploy the Cisco Nexus 1000v for KVM Virtual Supervisor Module (VSM) and associated Virtual Ethernet Module (VEM) for a robust and policy aware virtual network control and data plane in OpenStack Neutron.
  7. Also a big Post-Deployment step: Implementing the OpenStack Object Storage Service (Swift) in a redundant and highly-available fashion, to take advantage of the storage provided by the NetApp E-Series E5560 through Dynamic Disk Pool (DDP) technology:

    • Configure for, access, and format for use iSCSI LUNs exposed to OpenStack Controller systems for Swift services.
    • Setup replication with a lowered count of two between the OpenStack Controller systems.
    • Configure Swift Account, Container, and Object Storage services
    • Setup the Swift Proxy service and build Swift rings throughout the cluster.
    • Configure HAProxy for Swift Proxy service High Availability, and add resources to the Pacemaker cluster to be in line with other HA services in OpenStack.

Finally, instructions are also provided to create and boot instances (virtual machines) in the newly resulting OpenStack Cloud so you can get IT resources up and running workloads quickly.

I hope this reference architecture and associated solution documentation is useful for you. Please feel free to leave comments in the section below!

For more info

To learn more about building a business-critical cloud for your organization, contact your NetApp, Cisco, or Red Hat sales representative or visit the following links: