udev & network interface order

_______________________________________________
Xen-users mailing list
Xen-users <at> lists.xensource.com
http://lists.xensource.com/xen-users

Re: udev & network interface order

Peter Fastré wrote:
> Hello
>
> I finally succeeded in creating my Slackware 11.0 domain0 with Xen 
> 3.0.4, with multiple network interfaces (thanks folks). Dom0 has eth0, 
> eth1, eth2 and also xenbr0, 1, 2. All domU's also have eth0/eth1/eth2, 
> each bound to its own bridge.
>
> My config file:
>    kernel = "/boot/vmlinuz-2.6-xenU"
>    memory = 256
>    name = "dns02"
>    vif = 
>
['mac=00:16:3e:00:02:00,bridge=xenbr0','mac=00:16:3e:00:02:01,bridge=xenbr1','mac=00:16:3e:00:02:02,bridge=xenbr2'] 
>
>    disk = 
> ['phy:/dev/vg/dns02_root,sda1,w','phy:/dev/vg/dns02_swap,sda2,w']
>    root = "/dev/sda1 ro"
>
> This is ok, all interfaces are created in the domU.
>
> The problem is only, the order in which they are created is random! On 
> one domU the order is like I want it to be, on some other domU's is not.
>
> I have disabled udev on the domU, because I read it somewhere. So I 
> would think the standard linux rules apply, I was expecting that the 
> interfaces would be ordered based on their hardware addresses. Not.
>
> So I tried starting udev anyway, and putting some rules in 
> /etc/udev/rules.d/network_interfaces:
>    KERNEL=="eth*", SYSFS{address}=="00:16:3e:00:02:00", NAME="eth0"
>    KERNEL=="eth*", SYSFS{address}=="00:16:3e:00:02:01", NAME="eth1"
>    KERNEL=="eth*", SYSFS{address}=="00:16:3e:00:02:02", NAME="eth2"
>
> After that, it seems ok on this domU. But should I be doing this? I 
> read somewhere on the mailing list that I should disable udev on the 
> domU. Why?
Udev works fine for me in RHEL, CentOS, and Fedora Core.

> A sign that it's wrong what I'm doing, is the following message on 
> startup:
>    Starting udevd:  /sbin/udevd --daemon
>    udevd-event[1318]: rename_netif: error changing net interface name 
> eth1_rename to eth0: No such device
>
> What should I do? Keep using udev, and ignoring the error on startup? 
> Will it work reliably? Or am I doing something wrong? Is there another 
> way to order my network devices in the domU? Maybe without using udev.
Depends on the OS. For RedHat based systems, there's a poorly documented 
"HWADDR=[macaddress]" option in 
/etc/sysconfig/network-scripts/ifcfg-eth* to set these.

Possible to run Xen inside QEMU?

Hello,

to test a few setups I thought running Xen inside QEMU might be
possible.
But when I tried it, Xen fails to boot, when using "noreboot" I did not
find any error-message:
<http://picpaste.de/Q_Screenshot_1.png>

Is it impossble, oder did I something wrong? (The "normal" Kernel boot
fine)

regards

Johannes

Re: Feisty Fawn

On Sat, 2007-03-31 at 15:47 +0200, Cyril Gouget wrote:

> I tried to copy my old .hvm file and the /etc/xen/xend- config.sxp
> file but when I create the domU, I get only a black screen with a "_"
> on the upper left corner. (the images still run in edgy)

I'm with this problem too :(

Anyway, if you can resolve this, please update the WIKI (and let me
know :-p)

https://help.ubuntu.com/community/XenVirtualMachine/XenOnUbuntuFeisty

setting vncdisplay in config (w/ fix)

I'm using xen installed with Fedora Core 7 test 2.  When trying to
explicitly set the vncdisplay value I get this error:

  File "/usr/lib64/python2.5/site-packages/xen/xend/server/vfbif.py", line 70
    args += ["--vncport", "%d" % (5900 + config["vncdisplay"])]
TypeError: unsupported operand type(s) for +: 'int' and 'str'

I only get this for paravirtualized guests.

I don't know python, but I was able to fix this with a cast around
config["vncdisplay"].  The new line looks like this:

    args += ["--vncport", "%d" % (5900 + int(config["vncdisplay"]))]

Larry

Xen 3.0.3 and RHEL 3

Hi,
     I am trying full virt install of RHEL 3 Update 5 in xen-3.0.3
for X86_64 using
Virt-Manager , while installing through ISO  the VM hangs with the
following message

CALLIN, before  setup_local_APIC()
Calibrating delay loop...

i tried giving no acpi apci  kernel boot option it did not help, i am
using AMD-V enabled CPU

attached is the screen shot of the error message

Satya

_______________________________________________
Xen-users mailing list
Xen-users <at> lists.xensource.com
http://lists.xensource.com/xen-users

Re: Xen 3.0.3 and RHEL 3

satyaakam goswami wrote:
> Hi,
>     I am trying full virt install of RHEL 3 Update 5 in xen-3.0.3
> for X86_64 using
> Virt-Manager , while installing through ISO  the VM hangs with the
> following message
*WHY*? RHEL is up to RHEL 5, and even RHEL 4 was pretty old and getting 
seriously out of date. Attempting to continue to support such an old OS 
when RHEL 5 includes direct support for Xen, and when x86_64 was quite 
experimental under RHEL 3, is just begging for pain.

Even if RHEL 5 is out of your price range right now, CentOS 4 supports 
Xen well via add-in RPM's from Xensource, and CentOS 5 should be out 
this week. Both should operate better in the x86_64 world.

DomU time local vs. GMT

_______________________________________________
Xen-users mailing list
Xen-users <at> lists.xensource.com
http://lists.xensource.com/xen-users

are clock sync issues fixed in 3.0.4?

I rolled out my server just days before 3.0.4 was released, and I have 
the annoying clock drift problem (independent_wallclock set to 0, but 
domUs all 10+ minutes fast of dom0).

Does anyone know if this issue addressed/resolved in Xen 3.0.4?

Thanks,

-C-

Resolved: DomU time local vs. GMT

_______________________________________________
Xen-users mailing list
Xen-users <at> lists.xensource.com
http://lists.xensource.com/xen-users