Deployment Guide

Error! No text of specified style in document.
13 Scaling VCF Domains Beyond a Single MX7000 Chassis | Document ID
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. Since the Management Domain cluster
and Workload cluster are currently deployed to a single MX7000, we can treat them similarly. See the Dell
EMC PowerEdge MX SmartFabric Configuration and Troubleshooting Guide for the process of deploying the
Management Domain. The four Management Domain hosts are configured in a vSAN, contain NSX-T Edge
Nodes and are configured for vMotion. Workload Domain hosts may be different than the Management
Domain hosts so a different Template will be created for Management and Workload hosts.
Since redistributing these management hosts across more than one chassis may be desirable in the future,
the compute sleds were deployed using the “Deploy To Devices” method. When the hosts are later moved to
the second MX7000 chassis SmartFabric identifies the ports to which the compute sled is connected and
apply the appropriate network settings. The “Deploy To Devices” method also means that the server
Configuration Elements were already applied and will not be changed on the hosts.
Initially a single MX7000 was deployed with a pair of MX9116n FSEs in the A1 and A2 slots. The FSEs were
configured in a nonspanning tree Ethernet configuration that connected to an LACP-managed LAG on the
upstream network layer. While the MX9116n is capable of Layer 3 routing when configured in Full Switch
Mode it was configured using SmartFabric. In this case, the MX9116n is connected to the upstream network
at Layer 2 and allows the upstream network handle the routing.
Single MX700 with dual MX9116n
This deployment of two MX9116n FSEs represents the starting point for a recommended design. The two
MX9116n FSEs can be redistributed as additional chassis are added and MX7116n FEMs can be added to
expand the port count of these FSEs. There are several possible configurations that are supported by
different combinations of the FSEs and FEMs. See the latest Dell EMC PowerEdge MX SmartFabric
Configuration and Troubleshooting Guide. All these configurations are designed for scalability of
deployment, management, and network capacity. As per the guide the following restrictions and guidelines
are in place when building a Scalable Fabric: