Re: [Xen-users] Networking with more NIC's

["Peter Fastré" <peter.fastre@xxxxxxxxx>]

On 3/17/07, jez <jez@xxxxxxxxxx> wrote:

Re-read Peter. I also asked for the output of *ifconfig* before. Better
yet, please post the output of "ip addr list" before starting xend.

Sorry, I really appreciate your help, and I will try to give you  all the information you need.
To avoid messing up these mailing list, I've put the information in a text file:
    http://peter.lunatis.be/xen/info.txt
(please let me know if I should mail the full info.txt)

>
> > What is it that you actually want xen to do with these three cards?
>
> The server has  3 intel cards. For the moment, I just want to use two of
> them, one for the public addresses (195.x.x.x) which is different for each
> vm, and one for a private network which would be used to mirror some things
> between servers.
>

Are you saying you want each DomU to have two interfaces eth0 and eth1.
One of those interfaces, eth0, should be connected to a bridge that
contains Dom0's public nic. The other one, eth1, should be connected to a
bridge that contains Dom0s private nic?

That's exactly what I want.
But as you said in your other mail, for now I would be happy with a functioning eth0 only (on dom0 & domU's).

Yeah, I said this wasn't going to solve your problem. I was just throwing
in some extra advice for free.

I do appreciate your help. The big problem with Xen is (imho) that's is difficult to setup, because of the very limited documentation. But I've come this far (had some other trouble before) and now I don't want to give up!

> Hotplug -> do they mean udev? Udev is not changed compared to the normal
> installation, I think it's working normally as it should. I did disable it
> for the test domU's on my test machines (after I succesffully booted them).
> But on this server I can't boot any domain.
>

The scripts in /etc/xen/scripts are called via the udev system. When xen
is installed it puts a set of rules in /etc/udev/ in a file called
xen-backend.rules and then adds a symlink to that file in
/etc/udev/rules.d. At least that is how it's done in the Debian
installation. I don't know how you copied accross your xen system from
your test server, but you should probably check that there are some xen
rules somewhere under /etc/udev.

This is exactly how it's done on my system. There's a set of rules (xen-backend.rules) in the directory /etc/udev, and  the symlink under rules.d is there.
Contents of the file:
SUBSYSTEM=="xen-backend", KERNEL=="tap*", RUN+="/etc/xen/scripts/blktap $env{ACTION}"
SUBSYSTEM=="xen-backend", KERNEL=="vbd*", RUN+="/etc/xen/scripts/block $env{ACTION}"
SUBSYSTEM=="xen-backend", KERNEL=="vtpm*", RUN+="/etc/xen/scripts/vtpm $env{ACTION}"
SUBSYSTEM=="xen-backend", KERNEL=="vif*", ACTION="" RUN+="$env{script} online"
SUBSYSTEM=="xen-backend", KERNEL=="vif*", ACTION="" RUN+="$env{script} offline"
SUBSYSTEM=="xen-backend", ACTION="" RUN+="/etc/xen/scripts/xen-hotplug-cleanup"
KERNEL=="evtchn", NAME="xen/%k"
KERNEL=="blktap[0-9]*", NAME="xen/%k"

However, I really doubt this is your problem. TBH I really can't see how
the xen scripts could be expected to do the right thing with your
present configuration as it stands. If you can answer my question above
about what exactly you want to do with eth0, eth1, and eth2, then we
can try and get you a configuration that works.

 

> >>
> >> xenbr0    Link encap:Ethernet  HWaddr 00:00:00:00:00:00
> >>          UP BROADCAST RUNNING NOARP  MTU:1500  Metric:1
> >>          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
> >>          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
> >>          collisions:0 txqueuelen:0
> >>          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
> >>
> >> xenbr1    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
> >>          UP BROADCAST RUNNING NOARP  MTU:1500  Metric:1
> >>          RX packets:1184 errors:0 dropped:0 overruns:0 frame:0
> >>          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
> >>          collisions:0 txqueuelen:0
> >>          RX bytes:67438 (65.8 KiB)  TX bytes:0 (0.0 b)
> >>
<snip/>
> root@vm01:~# brctl show
> bridge name     bridge id               STP enabled     interfaces
> xenbr1          8000.feffffffffff       no              vif0.1
>                                                         peth0


Your ifconfig says you've got two bridges xenbr0 and xenbr1 but the
output of 'brctl show' says you only have one bridge. Can you double
check this, thanks.

That's ok now. As you can see in the text file, there's one bridge in the ifconfig xenbr1. But xenbr1, does this mean this is a bridge to eth1? Probably not, because it seems to work this way on my test machine.

Now I will try your suggestions in the other mail!

Regards

Peter

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