Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
46
Installation and Upgrades / webmail module in Core 3.1.8
« on: September 07, 2013, 09:30:08 am »
In the Roundcube Module, in main.inc.php, it reads:
// store spam messages in this mailbox
$rcmail_config['junk_mbox'] = 'Spam';
...
// display these folders separately in the mailbox list.
// these folders will also be displayed with localized names
$rcmail_config['default_imap_folders'] = array('INBOX', 'Drafts', 'Sent', 'Junk', 'Trash');
Instead, shouldn't this be either:
or:
OL uses Junk as default folder other common webmailers and clients (tb etc.) use Spam.
// store spam messages in this mailbox
$rcmail_config['junk_mbox'] = 'Spam';
...
// display these folders separately in the mailbox list.
// these folders will also be displayed with localized names
$rcmail_config['default_imap_folders'] = array('INBOX', 'Drafts', 'Sent', 'Junk', 'Trash');
Instead, shouldn't this be either:
Quote
// display these folders separately in the mailbox list.
// these folders will also be displayed with localized names
$rcmail_config['default_imap_folders'] = array('INBOX', 'Drafts', 'Sent', 'Spam', 'Trash'); //common
or:
Quote
// store spam messages in this mailbox
$rcmail_config['junk_mbox'] = 'Junk'; //windows OL
OL uses Junk as default folder other common webmailers and clients (tb etc.) use Spam.
47
Installation and Upgrades / Re: Remote desktop
« on: September 06, 2013, 01:26:40 pm »
I am confused.
Do you simply want the shares on your (external) windows client somewhere or a complete desktop environment?
Do you want to passthrough (the server) from home to your pc on work?
If you want remote desktop protocol (RDP) you may want to look at xrdp.
(You also want to think about "Thin clients" to the zentyal server)
hth Markus
Do you simply want the shares on your (external) windows client somewhere or a complete desktop environment?
Do you want to passthrough (the server) from home to your pc on work?
If you want remote desktop protocol (RDP) you may want to look at xrdp.
(You also want to think about "Thin clients" to the zentyal server)
hth Markus
48
Installation and Upgrades / Re: Remote desktop
« on: September 03, 2013, 09:29:53 pm »
I suggest remmina from a linux desktop to a tightvncserver on the zentyal box to go to the zentyal (console) desktop. You'll find some of the most wanted protocols.
You may want to have a look at remmina also, if you need RDP access from zentyal (console) to one of your windows clients. But it's not integrated in the zentyal desktop, would be a nice idea from manage computers for a service desk
In either case do a
hth Markus
You may want to have a look at remmina also, if you need RDP access from zentyal (console) to one of your windows clients. But it's not integrated in the zentyal desktop, would be a nice idea from manage computers for a service desk
Code: [Select]
sudo apt-get install remminaIf you want a RDP access from a windows to the zentyal server desktop, you may want to install ...Code: [Select]
sudo apt-get install xrdpIf you want VNC access from the windows client to the server you may want to set up on your zentyal box ...Code: [Select]
sudo apt-get install tightvncserver... and use a vnc client of your choice in windows, for example TightVNC.In either case do a
Code: [Select]
netstat -npa | grep -i "tcp.*:\*.*\(xrdp\|xvnc\)" and check your firewall for opening the door - avoid any of these without a tls tunnel (ssh/openvpn) from the outside in.hth Markus
49
Installation and Upgrades / Re: Issue with LDAP - Users and Contacts
« on: August 27, 2013, 08:59:34 am »
1+2 agreed!
50
Installation and Upgrades / Re: Issue with LDAP - Users and Contacts
« on: August 27, 2013, 06:53:30 am »
There are no Contacts by default in 3.1.7, but you understood right that there is the ability to create own OUs. I do not need it in the GUI, there are too less fields in the offered contacts scheme and I dont know who needs this little piece at all. I'd like to introduce a ou=Contacts... tree here for the thunderbird adressbook. Rather than syncing it I want to have one addressbook for all, offered from the server side.
51
Installation and Upgrades / Issue with LDAP - Users and Contacts
« on: August 26, 2013, 11:00:22 am »
Regarding version 3.1.7
I have done a ou=Office,ou=Contacts under top level tree.
I used phpldapadmin to export a contact entry from a ou=Company,ou=Domain Controllers, modified and imported it to ou=Office,ou=Contacts
When looking at phpldapadmin the entry is ok. In the Zentyal UI the entry is moved back to the former ou and deleted from ou=Office,ou=Contacts. After returning to phpldapadmin the entry has disappeared and is back in the old tree.
Is there something like a howto around how I can achive Thunderbird entries in a specific ou within the forest and/or moving users around within certain ou's (example a Sales Person from ou=sales to ou=managers)?
Thanks for any help...
Markus
I have done a ou=Office,ou=Contacts under top level tree.
I used phpldapadmin to export a contact entry from a ou=Company,ou=Domain Controllers, modified and imported it to ou=Office,ou=Contacts
When looking at phpldapadmin the entry is ok. In the Zentyal UI the entry is moved back to the former ou and deleted from ou=Office,ou=Contacts. After returning to phpldapadmin the entry has disappeared and is back in the old tree.
Is there something like a howto around how I can achive Thunderbird entries in a specific ou within the forest and/or moving users around within certain ou's (example a Sales Person from ou=sales to ou=managers)?
Thanks for any help...
Markus
52
Installation and Upgrades / Re: DHCP module does not honor computer groups
« on: August 25, 2013, 04:48:20 pm »To be honest, I didn't look closely at the resulting dhcp.conf file except for adding some DHCP option in mas file.
Using Zentyal GUI, it doesn't really make sense for me, most of the time, to closely look at conf files if I don't face any problem
Sorry, it's my 1st install, so I try to understand what's happening behind the scenes.
But I understand your position, I'm on the way to trust ... will be there soon
53
Installation and Upgrades / Re: DHCP module does not honor computer groups
« on: August 25, 2013, 08:18:36 am »
The remaining question is: Where do the empty blocks come from in dhcpd.conf:
group {
}
(These are between different computer groups)
Greets Markus
group {
}
(These are between different computer groups)
Greets Markus
54
Installation and Upgrades / [Solved] Re: DHCP module does not honor computer groups
« on: August 24, 2013, 10:59:28 pm »
I've forgotten, RTFM! Reading helps ...
Be my guest and have some coffee
Be my guest and have some coffee
55
Installation and Upgrades / DHCP module does not honor computer groups
« on: August 24, 2013, 10:10:20 am »
If you define computer groups within the network module (lets say network printers), then add devices - and give all information, as name, ip and mac address - the dhcp module will not generate them as static hosts within this block. How can I get to the goal?
Greets Markus
Greets Markus
56
Installation and Upgrades / [HOWTO] interface bonding using bond0 (also connected to bridge br0)
« on: August 24, 2013, 12:37:39 am »
In the 3.2 core it's easy to use bonding with the network.preservice, if you also need a bridge (example for kvm-qemu machines, connected to the outer world)
Prerequisites:
- Your interfaces support bonding (use ethtool to check)
- Configure the first bridge (br?) in zentyal network module and use the fixed address you need.
If you have modified the DFT variables, out of the box the code is desinged to work like ...
[lan or wan]
/ \
[eth0] [eth1]
\ /
[bond0]
|
[br?]
Warning: Dont use the following script without modification if you have more than one interface configured or virtual cards on your bridge in zentyal.
In the case you're still here, continue with ...
Then put a hook in /etc/zentyal/hooks/network.preservice with the following ...
Reconfiguring the zentyal dhcp module might be needed and also check your gateway/routing in the GUI
I had some trouble to get the routing back stable and have added a second (fallback) script as /etc/zentyal/hooks/network.postservice for the default route ...
You could try ifdown br0 and then ifup br0 to see wether things are working. Be near to the console (just in case) or have IPMI access and enjoy your bonded if's.
You may also want to edit /etc/zentyal/network.conf:
Find the line containing:
ifaces_to_ignore = sit,tun,tap,lo,...
and change it to:
ifaces_to_ignore = bond,sit,tun,tap,lo,...
to avoid sambas searching for an ip on the bond interface (see sambas log). The drawback is you wont have access to this interface in the firewall.
It's easy to adapt if you just need bonding but no bridge, do not make the change to /etc/zentyal/network.conf and move the address configuration part of br to the bond defs and then remove the rest br part from code, see the following example ...
In recent versions of Ubuntu the code is very straight, you simply copy the following script to /etc/zentyal/hooks/network.preservice ...
If you are using KVM/QEMU, for a powerful and fast way of bridging, and if you are using bonding without a (userspace) own bridge, you may want to look at macvtap device support for the kvm/qemu (modprobe vhost_net) which will give you a kernel bridge and full access from internet side if you use the virtio bridge mode. The xml part for qemu looks similar like ...
The drawback of this solution is, you wont ping your virtual host from bare metal host and may want to have a second (internal bridge) interface to access host from virtual host and vs. - or return to the first solution here. You then use the vhost-net to connect via a vnet interface on the bridge. You still can use a macvtap device, connected to the bond, to achive full speed on an other kvm machine. Some may think about "if I use Zentyal in a KVM and setup a second one in a KVM I can test changes before going to production mode"...
... adapt the settings for your interfaces as needed, suggestions are welcome.
Maybe someone could come along with a nice UI part for this purpose in the network module.
hth Markus
You may want to read https://www.kernel.org/doc/Documentation/networking/bonding.txt and
an example for a complex network config can be found here: https://www.stgraber.org/download/complex-interfaces
A mac-address generator for the bridge was found here http://www.miniwebtool.com/mac-address-generator/
Prerequisites:
- Your interfaces support bonding (use ethtool to check)
- Configure the first bridge (br?) in zentyal network module and use the fixed address you need.
If you have modified the DFT variables, out of the box the code is desinged to work like ...
[lan or wan]
/ \
[eth0] [eth1]
\ /
[bond0]
|
[br?]
Warning: Dont use the following script without modification if you have more than one interface configured or virtual cards on your bridge in zentyal.
In the case you're still here, continue with ...
Code: [Select]
sudo apt-get install ifenslave
sudo modprobe bonding
sudo grep -q bonding /etc/modules || echo bonding >>/etc/modules
Then put a hook in /etc/zentyal/hooks/network.preservice with the following ...
Code: [Select]
#!/bin/bash
# Revision: 0.57
# Date: 2013-09-27
# Author: Markus Neubauer
# License: GPLv2
# Zentyal Version: 3.2
# Description: Add bonding with bond0 using eth0/1 connected to a bridge on br?
# Use it as /etc/zentyal/hooks/network.preservice and make
# it executable
# Remarks: Placed in public for the zentyal community at:
# http://forum.zentyal.org/index.php/topic,17839.html
# This script can safely be removed when zentyal has official support for bonded interfaces
# Prerequisites:
# sudo apt-get install ifenslave
# comment next line out, if done so already in /etc/modules...
lsmod | grep -q "^bonding" || modprobe bonding
# uncomment next line to stop samba from searching an ip for bond0
# sed -i "s#ifaces_to_ignore = sit,tun#ifaces_to_ignore = bond,sit,tun#" /etc/zentyal/network.conf
# trunk interfaces
RAW_IF1="eth0" # modify this if using the new naming
RAW_IF2="eth1" # modify this if using the new naming
# ... extend to your needs and modify bond0 block if extended
# change "balance-alb" to your bonding mode if required
# Declare your own fallback variables, just in case something goes wrong
# adapt to your needs and put some meaningful values here
# you may safely want to delete the default definition for HWaddr, but set up your values for the rest
declare -r DFT_IF='0 inet static' # THIS IS br0, oimitting the "br"
declare -r DFT_IPAddress='192.168.X.X'
declare -r DFT_NetMask='255.255.255.X'
declare -r DFT_BroadCast='192.168.X.255'
declare -r DFT_GateWay='192.168.X.X'
declare -r DFT_HWaddr='XX:XX:XX:XX:XX:XX' # you may want to try http://www.miniwebtool.com/mac-address-generator/
if [ 'XX:XX:XX:XX:XX:XX' = "${DFT_HWaddr}" ]; then
echo "DOING NOTHING IN HOOK ${0} - NOT CONFIGURED on `date`" >>/etc/network/interfaces
exit 1
fi
# backup zentyal config (debugging)
# makes you find a /etc/network/.interfaces.zentyal-hook.*.last
# which is the zentyal's fromer config
TMP_FILE="/etc/network/.interfaces.zentyal-hook.${0##*/}.last"
SAVE_FILE="${TMP_FILE}-`date`"
cp /etc/network/interfaces "${SAVE_FILE}"
[ -L "${TMP_FILE}" ] && rm "${TMP_FILE}"
ln -s "${SAVE_FILE}" "${TMP_FILE}"
TMP_FILE=`tempfile`
# save zentyal network module config variables
grep -m1 -A5 "iface br" /etc/network/interfaces >${TMP_FILE}
TEMP_VAR=`grep -m1 'iface br' ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*iface br}
IF=${TEMP_VAR:-${DFT_IF}}
TEMP_VAR=`grep -m1 address ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*address }
IPAddress=${TEMP_VAR:-${DFT_IPAddress}}
TEMP_VAR=`grep -m1 netmask ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*netmask }
NetMask=${TEMP_VAR:-${DFT_NetMask}}
TEMP_VAR=`grep -m1 broadcast ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*broadcast }
BroadCast=${TEMP_VAR:-${DFT_BroadCast}}
TEMP_VAR=`grep -m1 gateway ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*gateway }
GateWay=${TEMP_VAR:-${DFT_GateWay}}
TEMP_VAR=`grep -m1 'hwaddress ether' ${TMP_FILE}`
if [ -n "${DFT_HWaddr}" ]; then
HWaddr=${DFT_HWaddr}
elif [ -f /sys/class/net/br${IF:0:1}/address ]; then
HWaddr=`cat /sys/class/net/br${IF:0:1}/address`
elif [ -n "${TEMP_VAR}" ]; then
HWaddr=${TEMP_VAR##*hwaddress ether }
fi
rm ${TMP_FILE}
# start a new config file with overwritting the previous one
echo "# replaced by zentyal hook ${0}
# on $(date)
auto lo
iface lo inet loopback
auto ${RAW_IF1}
iface ${RAW_IF1} inet manual
bond-master bond0
auto ${RAW_IF2}
iface ${RAW_IF2} inet manual
bond-master bond0
# add more raw interface blocks if needed (and extend the following bond)
# leave the bonded device entries above the bond entry
# to avoid race conditions in certain environments
auto bond0
iface bond0 inet manual
bond-mode balance-alb # should be selectable in GUI
bond-miimon 100
bond-lcap-rate 1
bond-slaves ${RAW_IF1} ${RAW_IF2}
up /sbin/ifconfig ${RAW_IF1} up || /bin/true
up /sbin/ifconfig ${RAW_IF2} up || /bin/true
up /sbin/ifenslave bond0 ${RAW_IF1} ${RAW_IF2} || /bin/true
down /sbin/ifenslave -d bond0 ${RAW_IF1} ${RAW_IF2} || /bin/true
post-down /sbin/ifconfig ${RAW_IF2} down || /bin/true
post-down /sbin/ifconfig ${RAW_IF1} down || /bin/true
auto br${IF:0:1}
iface br${IF}" >/etc/network/interfaces
# restore zentyal network module config variables (append)
echo " address ${IPAddress}" >>/etc/network/interfaces
echo " netmask ${NetMask}" >>/etc/network/interfaces
echo " broadcast ${BroadCast}" >>/etc/network/interfaces
echo " gateway ${GateWay}" >>/etc/network/interfaces
[ -n "${HWaddr}" ] && echo " hwaddress ether ${HWaddr}" >>/etc/network/interfaces
# append the rest of your config
echo " bridge_ports bond0
bridge_fd 9
bridge_hello 2
bridge_maxage 12
bridge_stp off
pre-up ifup bond0
up /sbin/ifconfig ${RAW_IF1} up || /bin/true
up /sbin/ifconfig ${RAW_IF2} up || /bin/true
down /sbin/ifconfig ${RAW_IF1} down || /bin/true
down /sbin/ifconfig ${RAW_IF2} down || /bin/true
post-down ifdown bond0
# in the case you are using virtual hosts,
# you'll have to add the defines here, too!
#auto br${IF:0:1}:yourvhost
#iface br${IF:0:1}:yourvhost inet static
# address <your_other_ip>
# netmask <your_mask>
# broadcast <your_bcast>
## eof" >>/etc/network/interfaces
exit 0
Make the script executable and test it by calling it yourself, review result in /etc/network/interfaces. Even if not all defs are necessary, i've tried to make the code failsafe...Reconfiguring the zentyal dhcp module might be needed and also check your gateway/routing in the GUI
I had some trouble to get the routing back stable and have added a second (fallback) script as /etc/zentyal/hooks/network.postservice for the default route ...
Code: [Select]
#!/bin/sh
# change <your_gateway> to a meaningful ip and uncomment next line
#route -n | grep -q "^0.0.0.0" || /sbin/route add default gw <your_gateway>
(make it executable and test)You could try ifdown br0 and then ifup br0 to see wether things are working. Be near to the console (just in case) or have IPMI access and enjoy your bonded if's.
You may also want to edit /etc/zentyal/network.conf:
Find the line containing:
ifaces_to_ignore = sit,tun,tap,lo,...
and change it to:
ifaces_to_ignore = bond,sit,tun,tap,lo,...
to avoid sambas searching for an ip on the bond interface (see sambas log). The drawback is you wont have access to this interface in the firewall.
It's easy to adapt if you just need bonding but no bridge, do not make the change to /etc/zentyal/network.conf and move the address configuration part of br to the bond defs and then remove the rest br part from code, see the following example ...
Code: [Select]
#!/bin/bash
# Revision: 0.58
# Date: 2013-10-04
# Author: Markus Neubauer
# License: GPLv2
# Zentyal Version: 3.2
# Description: Add bonding with bond0 using eth0/1
# Use it as /etc/zentyal/hooks/network.preservice and make
# it executable
# Remarks: Placed in public for the zentyal community at:
# http://forum.zentyal.org/index.php/topic,17839.html
# This script can safely be removed when zentyal has official support for bonded interfaces
# Prerequisites:
# sudo apt-get install ifenslave
# comment next line out, if done so already in /etc/modules...
lsmod | grep -q "^bonding" || modprobe bonding
# trunk interfaces
RAW_IF1="eth0" # modify this if using the new naming
RAW_IF2="eth1" # modify this if using the new naming
# ... extend to your needs and modify bond0 block if extended
# change "balance-alb" to your bonding mode if required
# Declare your own fallback variables, just in case something goes wrong
# adapt to your needs and put some meaningful values here
# you may safely want to delete the default definition for HWaddr, but set up your values for the rest
declare -r DFT_IF='0 inet static' # THIS IS bond0, oimitting the "bond"
declare -r DFT_IPAddress='192.168.X.X'
declare -r DFT_NetMask='255.255.255.X'
declare -r DFT_BroadCast='192.168.X.255'
declare -r DFT_GateWay='192.168.X.X'
# backup zentyal config (debugging)
# makes you find a /etc/network/.interfaces.zentyal-hook.*.last
# which is the zentyal's fromer config
TMP_FILE="/etc/network/.interfaces.zentyal-hook.${0##*/}.last"
SAVE_FILE="${TMP_FILE}-`date`"
cp /etc/network/interfaces "${SAVE_FILE}"
[ -L "${TMP_FILE}" ] && rm "${TMP_FILE}"
ln -s "${SAVE_FILE}" "${TMP_FILE}"
TMP_FILE=`tempfile`
# save zentyal network module config variables
grep -m1 -A5 "iface bond" /etc/network/interfaces >${TMP_FILE}
TEMP_VAR=`grep -m1 'iface bond' ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*iface bond}
IF=${TEMP_VAR:-${DFT_IF}}
TEMP_VAR=`grep -m1 address ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*address }
IPAddress=${TEMP_VAR:-${DFT_IPAddress}}
TEMP_VAR=`grep -m1 netmask ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*netmask }
NetMask=${TEMP_VAR:-${DFT_NetMask}}
TEMP_VAR=`grep -m1 broadcast ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*broadcast }
BroadCast=${TEMP_VAR:-${DFT_BroadCast}}
TEMP_VAR=`grep -m1 gateway ${TMP_FILE}`
TEMP_VAR=${TEMP_VAR##*gateway }
GateWay=${TEMP_VAR:-${DFT_GateWay}}
rm ${TMP_FILE}
# start a new config file with overwritting the previous one
echo "# replaced by zentyal hook ${0}
# on $(date)
auto lo
iface lo inet loopback
auto ${RAW_IF1}
iface ${RAW_IF1} inet manual
bond-master bond0
auto ${RAW_IF2}
iface ${RAW_IF2} inet manual
bond-master bond0
# add more raw interface blocks if needed (and extend the following bond)
# leave the bonded device entries above the bond entry
# to avoid race conditions in certain environments
auto bond${IF:0:1}
iface bond${IF}" >/etc/network/interfaces
# restore zentyal network module config variables (append)
echo " address ${IPAddress}" >>/etc/network/interfaces
echo " netmask ${NetMask}" >>/etc/network/interfaces
echo " broadcast ${BroadCast}" >>/etc/network/interfaces
echo " gateway ${GateWay}" >>/etc/network/interfaces
# append the rest of your config
echo " bond-mode balance-alb # should be selectable in GUI
bond-miimon 100
bond-lcap-rate 1
bond-slaves ${RAW_IF1} ${RAW_IF2}
up /sbin/ifconfig ${RAW_IF1} up || /bin/true
up /sbin/ifconfig ${RAW_IF2} up || /bin/true
up /sbin/ifenslave bond0 ${RAW_IF1} ${RAW_IF2} || /bin/true
down /sbin/ifenslave -d bond0 ${RAW_IF1} ${RAW_IF2} || /bin/true
post-down /sbin/ifconfig ${RAW_IF2} down || /bin/true
post-down /sbin/ifconfig ${RAW_IF1} down || /bin/true
# in the case you are using virtual hosts,
# you'll have to add the defines here, too!
#auto bond${IF:0:1}:yourvhost
#iface bond${IF:0:1}:yourvhost inet static
# address <your_other_ip>
# netmask <your_mask>
# broadcast <your_bcast>
## eof" >>/etc/network/interfaces
exit 0
In the case your are using DHCP you should remove the addressing part and dont save the former setup in tmp which will make the script very short In recent versions of Ubuntu the code is very straight, you simply copy the following script to /etc/zentyal/hooks/network.preservice ...
Code: [Select]
#!/bin/bash
# Revision: 0.59
# Date: 2013-10-04
# Author: Markus Neubauer
# License: GPLv2
# Zentyal Version: 3.2
# Description: Add bonding with bond0 using eth0/1 on DHCP
# Use it as /etc/zentyal/hooks/network.preservice and make
# it executable
# Remarks: Placed in public for the zentyal community at:
# http://forum.zentyal.org/index.php/topic,17839.html
# This script can safely be removed when zentyal has official support for bonded interfaces
# Prerequisites:
# sudo apt-get install ifenslave
# comment next line out, if done so already in /etc/modules...
lsmod | grep -q "^bonding" || modprobe bonding
# change "balance-alb" to your bonding mode if required
# start a new config file with overwritting the previous one
echo "# replaced by zentyal hook ${0}
# on $(date)
auto lo
iface lo inet loopback
# raw network interface 1
auto eth0
iface eth0 inet manual
bond-master bond0
# raw network interface 2
auto eth1
iface eth1 inet manual
bond-master bond0
auto bond0
iface bond0 inet dhcp
bond-mode balance-alb
bond-miimon 100
bond-lcap-rate 1
bond-slaves eth0 eth1
iface bond0 inet6 auto
# in the case you are using virtual hosts,
# you'll have to add the defines here, too!
#auto bond0:yourvhost
#iface bond0:yourvhost inet static
# address <your_other_ip>
# netmask <your_mask>
# broadcast <your_bcast>
## eof" >/etc/network/interfaces
If you are using KVM/QEMU, for a powerful and fast way of bridging, and if you are using bonding without a (userspace) own bridge, you may want to look at macvtap device support for the kvm/qemu (modprobe vhost_net) which will give you a kernel bridge and full access from internet side if you use the virtio bridge mode. The xml part for qemu looks similar like ...
Code: [Select]
<interface type='direct'>
<mac address='<your mac address>'/>
<source dev='bond0' mode='bridge'/>
<model type='virtio'/>
<address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/>
</interface>
... which gives you the qemu parameter part -netdev tap,id=hostnet0,vhost=on -device virtio-net-pci,netdev=hostnet0,id=net0,mac=<your mac address>,bus=pci.0,addr=0x4The drawback of this solution is, you wont ping your virtual host from bare metal host and may want to have a second (internal bridge) interface to access host from virtual host and vs. - or return to the first solution here. You then use the vhost-net to connect via a vnet interface on the bridge. You still can use a macvtap device, connected to the bond, to achive full speed on an other kvm machine. Some may think about "if I use Zentyal in a KVM and setup a second one in a KVM I can test changes before going to production mode"...
... adapt the settings for your interfaces as needed, suggestions are welcome.
Maybe someone could come along with a nice UI part for this purpose in the network module.
hth Markus
You may want to read https://www.kernel.org/doc/Documentation/networking/bonding.txt and
an example for a complex network config can be found here: https://www.stgraber.org/download/complex-interfaces
A mac-address generator for the bridge was found here http://www.miniwebtool.com/mac-address-generator/