mirror of
https://github.com/Ponce/slackbuilds
synced 2024-11-29 13:00:32 +01:00
444996d10d
Signed-off-by: Niels Horn <niels.horn@slackbuilds.org>
39 lines
1.7 KiB
Text
39 lines
1.7 KiB
Text
Here are some scripts that I've written for use with Xen at my site.
|
|
In order to use these scripts with your Xen installation you'll need to copy
|
|
both the network-openvswitch and vif-openvswitch files to your
|
|
/etc/xen/scripts directory. You can instruct Xen to use these scripts by
|
|
editing your /etc/xen/xend-config.sxp file and specifying these scripts as
|
|
the default network-script and vif-script.
|
|
|
|
For example, here are the entries in my xend-config.sxp file:
|
|
|
|
(network-script 'network-openvswitch netdev=eth2 bridge=ovs0')
|
|
|
|
(vif-script 'vif-openvswitch bridge=ovs0')
|
|
|
|
If your network interface card and attached network switch support VLAN
|
|
tagged traffic, you can place virtual machines within a seperate VLAN by
|
|
appending a '.' and the VLAN tag number you wish the domain to use. For
|
|
example, to have all domains default to VLAN 2 you can do the following:
|
|
|
|
(vif-script 'vif-openvswitch bridge=ovs0.2')
|
|
|
|
You can also specify tagged traffic in the domain configuration file.
|
|
|
|
If you are hosting a hardware virtualized domain understand that the
|
|
/etc/xen/scripts/qemu-ifup script is run instead of the vif-script specified
|
|
in the xend-config.sxp file. You'll need to edit this file to add the port
|
|
to the vswitch instead of using the brctl (unless of course you are using
|
|
the appropriate kernel module to control the vswitch using brctl.) Here is
|
|
a snippet from my qemu-ifup to handle hardware virtualized ports:
|
|
|
|
if lsmod | grep -c openvswitch_mod 1> /dev/null && ! lsmod | grep -c brcompat_mod 1> /dev/null
|
|
then
|
|
ovs-vsctl -- --may-exist add-port $bridge $1
|
|
else
|
|
brctl addif $bridge $1 || true
|
|
fi
|
|
|
|
This doesn't handle tagged traffic. I'll leave that as an exercise for you.
|
|
|
|
Enjoy.
|