If you plan to use KVM as a transport node or as a host for NSX Manager guest VM, but you do not already have a KVM setup, you can use the procedure described here.
Procedure
- (Only RHEL) Open the /etc/yum.conf file.
- Search for the line exclude.
- Add the line
"kernel* redhat-release*"
to configure YUM to avoid any unsupported RHEL upgrades.exclude=[existing list] kernel* redhat-release*
If you plan to run NSX-T Data Center Container Plug-in, which has specific compatibility requirements, exclude the container-related modules as well.
The supported RHEL versions are 7.7 and 8.2.exclude=[existing list] kernel* redhat-release* kubelet-* kubeadm-* kubectl-* docker-*
- Install KVM and bridge utilities.
Linux Distribution Commands Ubuntu apt-get install -y qemu-kvm libvirt-bin ubuntu-vm-builder bridge-utils virtinst virt-manager virt-viewer libguestfs-tools
RHEL or CentOS Linux yum groupinstall "Virtualization Hypervisor" yum groupinstall "Virtualization Client" yum groupinstall "Virtualization Platform" yum groupinstall "Virtualization Tools"
SUSE Linux Enterprise Server Start YaSt and select Virtualization > Install Hypervisor and Tools. YaSt allows you to automatically enable and configure the network bridge.
- For NSX manager to automatically install NSX software packages on KVM host, prepare the network configuration of the uplink/data interface.
The KVM host can have multiple network interfaces. For the network interface that you plan to provide as an uplink interface (data interface) for NSX-T purposes, it is important to have network configuration files correctly populated. NSX-T looks at these network configuration files to create NSX-T specific network devices. On Ubuntu, populate /etc/network/interfaces file. On RHEL, CentOS, or SUSE, populate the /etc/sysconfig/network-scripts/ifcfg-$uplinkdevice file.
In the following examples, interface "ens32" is the uplink device (data interface). Depending on your deployment environment, this interface can use DHCP or static IP settings.
Note: Interface names might vary in different environments.Important: For Ubuntu, all network configurations must be specified in /etc/network/interfaces. Do not create individual network configuration files such as /etc/network/ifcfg-eth1, which can lead to failure of transport node creation.Linux Distribution Network Configuration Ubuntu Edit /etc/network/interfaces:
auto eth0 iface eth0 inet manual auto ens32 iface ens32 inet manual
RHEL or CentOS Linux Edit /etc/sysconfig/network-scripts/ifcfg-ens32:
DEVICE="ens32" TYPE="Ethernet" NAME="ens32" UUID="<something>" BOOTPROTO="none" HWADDR="<something>" ONBOOT="yes" NM_CONTROLLED="no"
SUSE Linux Enterprise Server If a SLES host already exists, verify that data interfaces is already configured on the host.
If you do not have a pre-configured SLES host, see the reference configuration for the management and data interface.
Edit /etc/sysconfig/network/ifcfg-ens32:
DEVICE="ens32" NAME="ens32" UUID="<UUID>" BOOTPROTO="none" LLADDR="<HWADDR>" STARTMODE="yes"
- Restart networking service systemctl restart network or reboot the Linux server for the networking changes take effect.
- After the KVM host is configured as a transport node, the bridge interface 'nsx-vtep0.0' is automatically created by NSX-T.
In Ubuntu, the /etc/network/interfaces file has entries such as the following:
iface nsx-vtep0.0 inet static pre-up ip addr flush dev nsx-vtep0.0 address <IP_pool_address> netmask <subnet_mask> mtu 1600 down ifconfig nsx-vtep0.0 down up ifconfig nsx-vtep0.0 up
In RHEL, the host NSX agent (nsxa) creates a configuration file named ifcfg-nsx-vtep0.0 that has entries such as the following:
DEVICE=nsx-vtep0.0 BOOTPROTO=static NETMASK=<IP address> IPADDR=<subnet mask> MTU=1600 ONBOOT=yes USERCTL=no NM_CONTROLLED=no
In SUSE,
DEVICE=nsx-vtep0.0 BOOTPROTO=static NETMASK=255.255.255.0 IPADDR=192.168.13.119 MACADDR=ae:9d:b7:ca:20:4a MTU=1600 USERCTL=no STARTMODE=auto
- Configure the syslog rotation policy as time-based instead of size-based policy. With a size-based syslog rotation policy, the log files generated might be of very large sizes.