
VMware ESXi Setup
The lab runs extremely well under ESXi; in my lab the basic vSphere cluster consumes a
peak of 4.5GB of RAM during build and about 40GB of disk space, although 220GB is
allocated. A higher performance disk system also reduces the lab build time as this is mainly
disk IOPS constrained.
The lab will usually use a portgroup on an Internal Only Standard vSwitch, i.e. one with no
physical NICs attached. The default portgroup name is Lab_Local and this portgroup must
be setup to allow Promiscuous mode and must be set to VLAN ID 4095. Your ESXi server
should also have a VMkernel port on this network, IP 192.168.199.99. This will allow access
to the Build share on the NAS VM.
The router VM also connects to your main network, the default configuration calls this
network External.
To enable the nested ESXi servers to run 64bit VMs your outer ESXi server needs to pass the
Intel VT or AMD V feature into the VMs. Open a command prompt on your ESXi server,
either through SSH or on the local console, and type the following command
Echo 'vhv.allow = "TRUE"' >> /etc/vmware/config
When you come to populate the build share and connect to the built VMs you will use the
Router VM to provide access into the lab network, as discussed in the accessing the lab
section.
Deploy on vSphere – vApp
If you have vCenter in your lab environment then you will be able to deploy the multi-VM
OVF. This requires either a DRS cluster or a standalone ESXi server, vApps cannot be
deployed to a cluster that has HA and not DRS.
Import the OVF, give the vApp a unique name and select the required datastore and
portgroups. The Lab_Local portgroup should map to the portgroup you created above and
the External portgroup should map to you normal production network from which you
access the vSphere environment. The router VM will consume one DHCP IP address from
the external network.
Deploy on ESXi – Single VM
Komentáře k této Příručce