vn.pod.1

来自「xen 3.2.2 源码」· 1 代码 · 共 177 行

=head1 NAME

vn - Vnet (virtual networking) management utility.

=head1 SYNOPSIS

vn <command> [args]

=head1 DESCRIPTION

The B<vn> utility manages vnets, virtual networks for virtual machines.
Before using vnets, the vnet kernel module must be installed or
the user-space daemon vnetd must be running. Using the kernel module is recommended,
see the B<insmod> command below.

A vnet is a virtual network that behaves like a private LAN, transporting
Ethernet frames. Each vnet is identified by a 128-bit vnet id and
has a network device that interfaces to it. Ethernet packets written
to the device are encapsulated and sent to the network.
Received vnet packets are decapsulated and delivered from the device
corresponding to their vnet id. The default encapsulation uses UDP on port 1798.

Usually each vnet device is enslaved to a corresponding bridge, and virtual
machine interfaces are attached to vnets by enslaving them to the bridge.
Each vnet behaves like a private LAN: traffic on one vnet is not visible
on other vnets, and interfaces on a vnet cannot see traffic on the
physical network. 

Vnets can be connected together into larger networks
by direct bridging or packet forwarding, or by using multihomed vms
with interfaces on several vnets, or vnets and the physical network.
As vnet interfaces are discovered dynamically, vnet connectivity is maintained
if a vm using a vnet is migrated from one physical machine to another.

In the commands vnet ids can be given in two forms. Long form, as 8 4-digit hex fields
separated by colons, for example 0000:0000:0000:0000:0000:0000:0000:0004, and
short form as a hex field, for example 0004 or 4. The short form is the same as the
long form with the first 7 fields zero. Vnet id 0000:0000:0000:0000:0000:0000:0000:0001
is reserved for the physical network and has no vnet device.

Vnets use multicast to discover the location of virtual interfaces, by default
using multicast group 224.10.0.1. If all the machines hosting vnets are on
the same subnet, or reachable by multicast, vnets will span all the machines
automatically. If some machines are not reachable by multicast you can configure
vnets to perform multicast forwarding using UDP. 

The vnet devices are fully-functional network devices, so you can add IP addresses
to them and test connectivity without any vms running.
For example, using vnif0004 on machines A and B:

        A> ifconfig vnif0004 10.0.0.11
        B> ifconfig vnif0004 10.0.0.12
        B> ping 10.0.0.11

If the vnet device is enslaved to a bridge you will have to add the IP address
to the bridge instead. Use C<brctl show> or C<vn vnets> to see if a vnet
device is on a bridge.

=over 4

=item B<insmod> I<[varp_mcaddr=ADDR]>

Insert the vnet kernel module, optionally supplying the multicast
address to use, default 224.10.0.1.

=item B<varp>

Print varp infrormation and varp cache.

=item B<vnets> [options]

Print the list of vnets (virtual networks). If a vnet device is on a bridge,
also shows the bridge and its bridged interfaces.

=over 4

=item B<-a | --all>

Also print the vifs on each vnet and varp information.

=item B<-l | --long>

Also print the ifconfig for the vnet devices.

=back

=item B<vnet-create> I<[options]> I<vnetid>

Create a vnet with the given id. The options are:

=over 4

=item B<-s | --security> I<level>

Security level, which can be one of I<none> for no security,
I<auth> for message authentication, and I<conf> for message
authentication and confidentiality. The default is no security.
Security is provided using IPSEC, but uses hard-wired keys.

=item B<-b | --bridge> I<bridgename>

Create a bridge for the vnet called I<bridgename> and enslave
the vnet device to it.

=item B<-v | --vnetif> I<vnetifname>

Use I<vnetifname> as the name for the vnet device. If this option
is not specified the default is to name the device vnifN where N
is the last field of the vnet id as 4 hex characters.
For example vnif0004. Network device names can be at
most 14 characters.

=back

=item B<vnet-delete> I<[options]> I<vnetid>

Delete the vnet with the given id. The vnet device goes away too.

=over 4

=item B<-b | --bridge>

If this option is specified, delete the bridge associated with the vnet.

=back

=item B<vifs>

Print the list of vifs (virtual interfaces).

=item B<vif-add> I<[-i|-interface]> I<vnet> I<vmac>

Add a vif to a vnet. Here I<vnet> is the vnet id and I<vmac>
is the vif's MAC address. Alternatively, I<vmac> can be the name of
a network device if the I<-i> or -I<--interface> flag is given.

It is not usually necessary to use B<vif-add> as vnets automatically
add vifs for the MAC addresses they see.

=item B<vif-delete> I<[-i|-interface]> I<vnet> I<vmac>

Delete a vif from a vnet. Here I<vnet> is the vnet id and I<vmac>
is the vif's MAC address. Alternatively, I<vmac> can be the name of
a network device if the I<-i> of -I<--interface> flag is given.

It is not usually necessary to use B<vif-delete> as vnets periodically
delete unused vifs.

=item B<peers>

Print the list of peer vnet machines to forward multicasts to, and accept
forwarded multicasts from.

=item B<peer-add> I<addr>

Add the peer with the given IP address or hostname.

=item B<peer-delete> I<addr>

Delete the peer with the given IP address or hostname.

=back

=head1 AUTHOR

The author of vn and vnets is Mike Wray of HP Labs. Please send problems, bugs,
enhancements requests etc. to [e-mail].

=head1 COPYRIGHT AND LICENSE

Copyright (C) 2006 Mike Wray <[e-mail]>.

This library is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version.