Document Scaling VCF Domains Beyond a Single MX7000 Chassis Using the MX7000 Scalable Fabric Architecture to expand VCF domains that are deployed to a single MX7000 chassis to multiple MX7000 chassis Abstract This technical paper outlines the considerations and process of expanding a VCF deployment to more than one MX7000 chassis leveraging the Scalable Fabric Architecture. This paper will discuss both the considerations of scalability of the initial deployment and the hardware used.
Revisions Revisions Date Description March 2021 Initial release Acknowledgments Author: Peter Giulietti The information in this publication is provided “as is.” Dell Inc. makes no representations or warranties of any kind with respect to the information in this publication, and specifically disclaims implied warranties of merchantability or fitness for a particular purpose. Use, copying, and distribution of any software described in this publication requires an applicable software license.
Table of contents Table of contents Revisions.............................................................................................................................................................................2 Acknowledgements .............................................................................................................................................................2 Table of contents .................................................................................................
Executive summary Executive summary Business IT departments are challenged to design and deploy private cloud infrastructure in an everexpanding environment. New compute, and storage resources are needed to process and store the increase of acquired data. IT departments can struggle to scale the compute, storage, and networking resources that are needed to meet their business needs.
Error! No text of specified style in document. 1 Planned Growth Scaling out any solution requires more than just the addition of hardware and software. Growth requires hardware that can be deployed quickly and consistently. In the case of a Hyper Converged Infrastructure, like VCF growth requires new clusters of compute, storage, and network resources that provide easy manageability and great flexibility. Deploying clusters has challenges and benefits.
Error! No text of specified style in document. 2 Dell EMC PowerEdge MX7000 With kinetic architecture and Agile management, the PowerEdge MX portfolio dynamically configures compute, storage, and fabric, increases team effectiveness, and accelerates operations. The responsive design delivers the innovation and longevity customers of all sizes need for their IT and digital business transformations. Key features of PowerEdge MX7000 include: • • • • • • 2.
Error! No text of specified style in document. MCM group cabling one, two, or three chassis Once the MX7000 chassis are assembled into an MCM group, network fabrics, deployment templates, networks, and other MX7000 management constructs can be used in the deployment and management of new and existing resources. 2.2 MX7000 Network Fabrics The MX7000 supports three I/O fabrics. The three fabrics are labeled: A, B, and C. Each fabric consists of a pair of I/O devices that provide redundant connectivity.
Error! No text of specified style in document. 3 Dell Scalable Fabric Architecture A Scalable Fabric spans multiple chassis and enables them to behave like a single chassis from a networking perspective. In contrast, an MCM group enables multiple chassis to be managed like a pool of chassis resources. A Scalable Fabric consists of two main components - one MX9116n FSE and one MX7116n FEM in each of the first two chassis, and additional pairs of MX7116n FEMs in the remaining chassis.
Error! No text of specified style in document. Networks also contain a Type value that pertains to the QoS settings for the network. There are 11 different network types available. Some examples are: Storage-iSCSI, Storage-FCoE, VM Migration and Hypervisor Management. The network type is selected at the time the network is created. Network types can be modified at any time. 3.2.2 Fabric Since each chassis can contain both an A and B pair of IOMs a chassis can have one or two fabrics.
Error! No text of specified style in document. “Ethernet” is used when there will be multiple, individual (nonaggregated) links to the upstream network. In this case Spanning Tree protocol is used to prevent loops in the Ethernet fabric that could have catastrophic results. The uplink only defines the behavior of the connection from the MX9116n IOMs to the upstream network. The uplink configuration has no impact on the server facing ports of the IOMs.
Error! No text of specified style in document. 4 Deployment Deploying host configuration parameters such as BIOS, iDRAC, RAID, NIC/CNA, and configuring the host facing switch ports is done through the Deployment process. Deployment involves Templates and the methods to apply the configuration data Templates can be deployed in two different ways.
Error! No text of specified style in document. The templates contain settings for the following categories: • • • • • • Local access configuration Location configuration Power configuration Chassis network configuration Slot configuration Setup configuration Templates are created by building templates from reference servers that have been configured, using one of the prebuilt templates in the OME-Modular interface or by importing a template that was exported from another system.
Error! No text of specified style in document. 5 Initial VCF Cluster Deployment When deploying the initial VCF Management Domain and Workload Domain on the first MX7000 consider growth beyond a single chassis. The choice of the Scalable Fabric Architecture is the fastest and most reliable path to a successful expansion. VCF Management and Workload Domains are deployed as clusters.
Error! No text of specified style in document. • • • • All MX7000 chassis in the same Scalable Fabric must be in the same Multichassis Group. Mixing IOM types in the same Scalable Fabric (for example, MX9116n FSE in fabric slot A1 and MX5108n in fabric slot A2) is not supported. All participating MX9116n FSEs and MX7116n FEMs must be in MX7000 chassis that are part of the same MCM group. For more information, see the Dell EMC OpenManage Enterprise-Modular Edition Version 1.20.
Error! No text of specified style in document. 6 Expand a VCF Cluster Expanding a cluster beyond a single MX7000 chassis involves. • • • • Shut down any cluster hosts that will be relocated Expanding the MCM group Reconfiguring the network fabric Bring the hosts and VMs back up Since the Management Domain cluster and Workload cluster are currently deployed to a single MX7000, you can treat them the same way.
Error! No text of specified style in document. Adding a member chassis From the Modify Group Members window (on the lead chassis) any other chassis that is properly cabled will appear in the Available Chassis pane. 1. Select the available chassis 2. Click Add Chassis. 3. Click Finish to complete the process and add the chassis to the MCM group.
Error! No text of specified style in document. Adding Member Chassis to MCM Group 6.3 Reconfiguring the Network Fabric Reconfiguration of the network fabric from one to two chassis is carried out by redistributing the MX9116n FSEs across the two chassis. The only invasive reconfiguration of the network fabric occurs when the fabric is expanded from a single chassis to more than one chassis. Adding additional chassis beyond the second chassis has no impact on workloads already running.
Error! No text of specified style in document. Scalable Fabric I/O Module Placement While there are other supported topologies this topology provides complete redundancy. See the Dell EMC PowerEdge MX SmartFabric Configuration and Troubleshooting Guide for more information about supported topologies. 6.3.
Error! No text of specified style in document. 6.3.2 Resulting Port Mappings Though the physical compute sleds were not moved the ports to which the NICs connect will change. The A2 NIC ports on the compute sled installed in slot 1 of chassis one, before the second chassis is added are shown in figure 9 below.
Error! No text of specified style in document. If, for example, you moved management domain hosts 3 and 4 to slots 1 and 2 of the new chassis the OMEM software would reconfigure the ports that are connected to new locations of those servers. 6.4 Restart Hosts and VMs Now that all the NIC ports and their corresponding FSE and or FEM ports have been reconfigured based on their location any hosts that were shut down can be brought back up.
Error! No text of specified style in document. 7 Conclusion The Dell EMC PowerEdge MX7000 coupled with the scalable fabric architecture provides an exceptionally flexible and reliable method for scaling out any VCF deployment. SmartFabric and deployment templates allow compute sleds to be moved between slots and chassis and maintain their specified host and network characteristics.