Deployment Guide

20 ScaleIO/VxFlex OS IP Fabric Best Practice and Deployment Guide with OS10EE | version 1.0
3.3.2 Create clusters and add hosts
When a host is added to a cluster, the host's resources become part of the cluster's resources. The cluster
manages the resources of all hosts within it. This section shows how to create a cluster.
All ESXi hosts are added to the cluster. The cluster name in this example is for identification purposes only.
To add clusters to the datacenter, complete the following steps:
1. On the web-client Home screen, select Hosts and Clusters.
2. In the Navigator pane, right-click the datacenter object and select New Cluster.
3. Name the cluster. For this example, the cluster is named atx01-w02-ScaleIO.
4. Leave DRS, vSphere HA, EVC and Virtual SAN at their default settings (Off/Disabled). Click OK.
Note: vSphere DRS, HA, and EVC cluster features are outside the scope of this guide. For more information
on these features, see the VMware vSphere 6.5 documentation.
In the Navigator pane, drag and drop ESXi hosts into the cluster.
When complete, each cluster ( ) should contain its assigned hosts ( ) as shown in Figure 11.
Clusters and hosts after initial configuration
3.3.3 Deploy vSphere distributed switch
A single vSphere distributed switch (VDS) is created in vCenter and deployed to each host. The VDS contains
port groups named management, ScaleIO-data01, ScaleIO-management, and vmotion. The management
port group supports management traffic to the ESXi hosts. The ScaleIO-data01 port group supports all
VxFlex OS traffic to include SDS-SDC and MDM traffic. The ScaleIO-management port group support
management traffic to VxFlex OS. Finally, the vmotion port group supports all vMotion traffic.
The VDS is assigned four vmnics from each host. Each host from the VxFlex OS cluster combine vmnic0 and
vmnic5 into an LACP-enabled port channel, lag 1. This lag is mapped to the management, ScaleIO-
management, and vmotion port groups. Each host from the VxFlex OS cluster combines vmnic1 and
vmnic4 into an LACP-enabled port channel, lag 2. This lag is mapped to the ScaleIO-data01 port group. In
this configuration, compute, or application traffic is forwarded over lag1, while storage traffic is forwarded over
lag2.