[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Bonding...again



Below is a working configuration in the form of ifcfg files, Xen network scripts, /etc/modprobe.conf, and /etc/rc.d/rc.local.  This is bonding without VLANs on a CentOS 5.2 x86-64 box.

ifcfg-eth0:
DEVICE=eth0
BOOTPROTO=none
HWADDR=00:A2:19:11:D1:58
>MASTER=bond0
SLAVE=yes

ifcfg-eth1:
DEVICE=eth1
BOOTPROTO=none
HWADDR=00:A2:19:11:D1:59
>MASTER=bond0
SLAVE=yes

ifcfg-bond0:
DEVICE=bond0
BOOTPROTO=none
>TYPE=ethernet
BONDING_OPTS="mode=4 miimon=100"

ifcfg-bond0:0:
DEVICE=bond0:0
BOOTPROTO=static
>TYPE=ethernet
IPADDR=10.1.1.5
NETMASK=255.255.255.0
DNS1=10.1.1.8
DNS2=10.1.1.9
DOMAIN="domain.com domain"
# Gateway is set in /etc/rc.d/rc.local

/etc/xen/xend-config.sxp (snippet):
(network-script network-multi-bridge)

/etc/xen/scripts/network-multi-bridge (this is actually a relic; I'm not sure why we're still using it):
#!/bin/sh
# network-multi-bridge
# Exit if anything goes wrong
set -e
# First arg is operation.
OP=$1
shift
script=/etc/xen/scripts/network-bridge
case ${OP} in
start)
$script start vifnum=0 bridge=xenbr0 netdev=bond0
;;
stop)
$script stop vifnum=0 bridge=xenbr0 netdev=bond0
;;
status)
$script status vifnum=0 bridge=xenbr0 netdev=bond0
;;
*)
echo 'Unknown command: ' ${OP}
echo 'Valid commands are: start, stop, status'
exit 1
esac

/etc/modprobe.conf (snippet):
alias bond0 bonding

/etc/rc.d/rc.local (snippet):
# Default gateway doesn't get set properly after the xend startup, so we'll set it here
route add -net default gw 10.1.1.1





On Fri, Aug 7, 2009 at 4:33 AM, Christian Fischer <Christian.Fischer@xxxxxxxxxxxxxxxxxxx> wrote:
On Thursday 06 August 2009, Adam Wead wrote:
> Thanks for the help...
>
> I've tried this method, as well as Christian's method which does not
> use a VLAN.  In both cases, I can get all interfaces to come up but
> can't get the guests to make network connections to the outside.  I
> should also note that under RedHat EL5.3, if you don't use a VLAN and
> bridge directly off of the bond, you run into this problem:
>
> https://bugzilla.redhat.com/show_bug.cgi?id=463014

Adam,

this is a simple setup you can do without any RHEL network init scripts.

I suggest writing your own with simple ip, brctl and ifenslave commands.
This is IMHO the shortest way.

Christian

>
> Which has been discussed on this list before.
>
> I think I'm getting hung up on the routing.  If we use your VLAN
> example, I configure bond0 with the public IP for dom0 (yes?), then I
> configure the VLAN with it's own network like 10.0.X.X or 192.168.X.X,
> then I create a route from the VLAN to bond0... yes?  If I understand
> bridging correctly, I won't need to add any routes from VLAN to
> bridge...
>
> Let me know if I've got that straight or not...
>
> Thanks again for the help...
>
> ...adam
>
> On Aug 5, 2009, at 12:18 AM, Jonathan Bayles wrote:
> > Take a look at these configs and the brctl command...(XXXX = vlan tag)
> >
> > [root@pnode006 /]# cat /etc/sysconfig/network-scripts/ifcfg-eth0.XXXX
> > DEVICE=eth0.XXXX
> > BOOTPROTO=static
> > > > > VLAN=yes
> > BRIDGE=brXXXX
> >
> > [root@pnode006 /]# cat /etc/sysconfig/network-scripts/ifcfg-brXXXX
> > DEVICE=brXXXX
> > TYPE=Bridge
> > BOOTPROTO=static
> > > > > DELAY=0
> > STP=off
> >
> > [root@pnode006 /]# brctl show
> > bridge name     bridge id               STP enabled     interfaces
> > brXXXX          8000.001ec9fd2360       no              vif3.0
> >                                                        eth0.XXXX vif
> > = [ "mac=YY:YY:YY:YY:YY:YY,bridge=brXXXX" ]
> >
> > Basically,
> >
> > Create a sub interface off your bond, configure it to do the
> > tagging, create a bridge for the tagged traffic, add the interface
> > onto a "port" on the bridge. Add the guest onto the bridge using a
> > vif.
> >
> > Subint <---- brctl bridge ----> guest vif
> >
> > -----Original Message-----
> > From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
> > [mailto:xen-users-bounces@xxxxxxxxxxxxxxxxxxx ] On Behalf Of Adam Wead
> > Sent: Tuesday, August 04, 2009 5:22 PM
> > To: Xen-users@xxxxxxxxxxxxxxxxxxx
> > Subject: [Xen-users] Bonding...again
> >
> > Hi all,
> >
> > I know this topic has been discussed a lot on this list, and I've been
> > going over some posts from months past trying to get things to work on
> > my end, but I need another appeal for help.  I'm setting up a dom0
> > under Red Hat EL5, with 2 NICs.  From looking over the old posts, it
> > looks like the best way to do this is to create bonds, vlans and
> > bridges under Red Hat and bypass Xen's initial network-bridge script.
> > I looked at what Fajar had said:
> >
> > http://lists.xensource.com/archives/html/xen-users/2009-05/msg00992.html
> >
> > And I tried duplicating this, mind you, without the patched ifup
> > script and setup something like this:
> >
> > - (eth1 + eth2) = bond0 with mode 2
> > - create VLAN bond0.2 to bond0
> > - attach bridge xenbr0 to bond0.2
> > - modified xend-config file: (vif-script 'vif-bridge bridge=xenbr0')
> >
> > This appears to work, but when I create guests and attach them to the
> > xenbr0, no dice.  My question is, is the above correct and where to
> > the IPs go?  Dom0 has it's own IP address, but what interface/bond/
> > VLAN does that go under?  Should the other items have IPs as well?
> >
> > My network kung fu is limited, so I apologize in advance if these are
> > stupid questions.
> >
> > many thanks,
> >
> > ...adam
> >
> >
> >
> >
> >
> > _______________________________________________
> > Xen-users mailing list
> > Xen-users@xxxxxxxxxxxxxxxxxxx
> > http://lists.xensource.com/xen-users



--
"Without music to decorate it, time is just a bunch of boring production
 deadlines or dates by which bills must be paid."
       --- Frank Vincent Zappa

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.