FC6/x86_64: missing veth*/vif*.* interfaces
by Axel Thimm
Hi,
I'm booting into a xen kernel (the latets Fedora update), but ifconfig
-a shows no interface pairs, and therefore xend and the network
scripts cannot setup any bridges etc.
Where/when are the interface pairs created? The system has been update
from FC4/x86_64 to FC5/x86_64 and FC6/x86_64, so maybe there is some
old bits I need to remove?
Thanks!
--
Axel.Thimm at ATrpms.net
17 years, 1 month
Xen domain0 and guest kernel match problem
by Xin Chen
Hi all,
In fc5xen installation guide, there is a troubleshoot section which
mentions that:
the kernel version for guest system must match with the domain0,
otherwise the installation may start normally and then go nowhere.
what does this mean? I got confusion that, does that mean I can't
install different version kernel guest on the domain0. for instance, I
use fc6 as domain 0, I can't install fc4 as guest system?? That doesn't
make any sense, does it?
can anyone help me ?
thanks!!
xin
17 years, 1 month
Running a presinstalled Windows XP on drive partition.
by NoisilySilent
Hi,
Did anyone succeed in running Windows preinstalled on its own partition?
Actually, I'm having my drive partitioned this way:
1) sda1 = ntfs (Windows install: directories \Windows and \Program
Files)
2) sda2 = ntfs (Windows install: directory \Document And Settings)
3) sda3 = ext3 (FC 6 with Xen)
4) sda4 = swap
The drive is full.
Partitions 1&2 were set long ago at work, and contain lots of private
documents and programs.
I wish I could run this installation of Windows within Fedora-Xen.
Is this possible? As all I can see is "how to install XP within Xen", not
running a real XP install!
Cheers,
Charles
17 years, 1 month
Xen and Cluster Manager (OpenAIS::TOTEM)
by Fabien MALFOY
Hi all
I experience problems with Xen while trying to build a two nodes cluster
with Cluster Manager. The Domain-0 host (called M0 here) is Fedora Core
6 and the Xen version is 3.0.3. Then, I created two virtual machines M1
and M2(FC6 too).
The matter is :
When i try to initialize the cluster using M1 and M2, these are not able
to start the service correctly. Cluster Manager depends on the new
cluster framework OpenAIS which uses the TOTEM protocol. Here is one
part of the log :
Mar 7 15:29:17 M1 openais[2220]: [CMAN ] CMAN 2.0.60 (built Jan 24 2007
15:30:39) started
Mar 7 15:29:17 M1 openais[2220]: [SYNC ] Not using a virtual synchrony
filter.
Mar 7 15:29:17 M1 openais[2220]: [MAIN ] AIS Executive Service: started
and ready to provide service.
Mar 7 15:29:18 M1 ccsd[2214]: Initial status:: Inquorate
*Mar 7 15:29:32 M1 openais[2220]: [TOTEM] The consensus timeout expired.*
*Mar 7 15:29:32 M1 openais[2220]: [TOTEM] entering GATHER state from 3.*
*Mar 7 15:29:47 M1 openais[2220]: [TOTEM] The consensus timeout expired.*
*Mar 7 15:29:47 M1 openais[2220]: [TOTEM] entering GATHER state from 3.*
*Mar 7 15:30:02 M1 openais[2220]: [TOTEM] The consensus timeout expired.*
*Mar 7 15:30:02 M1 openais[2220]: [TOTEM] entering GATHER state from 3.*
And this till the timeout...
As in this case I tried to build a cluster made of two virtualized
nodes, I wanted to be sure that the source of the problem was not my
configuration files or another reason of that type.
So I adapted my cluster.conf to replace M2 by M0 and then I installed
Cluster Manager on M0. When I started the service on M0 and M1, only the
one on M0 started correctly (understand "the non-virtualized machine")
and didn't receive any network data from M1.
For the tests, I disabled all network traffic filtering. So I deduce
that there's something strange between Xen and OpenAIS.
I would like to specify that I really tried all my knowledge to solve
this. I even thinked to disable the checksum offloading but no way.
Can anyone help me ? I also accept those who want to tell me they're in
the same lose :-D
--
Fabien MALFOY
+---------------------------------+
| Société StartX |
| 08.70.33.38.70 - 06.69.67.71.55 |
| http://startx.fr |
+---------------------------------+
17 years, 1 month
Problem about creating virtual machine in FC6
by taoj2@cs.rpi.edu
Hi,
I installed FC6 and I was trying to create a new virtual machine with the
virtual machine manager. I selected to use FC6 again as the guesting OS of
the new virtual machine.
Everytime after I finishing install of the guest OS, It gave me a prompt
to reboot the guest system. But After I click the "reboot" button, the
vitual machine gone. It's weird. I think the reboot shouldn't make the
guest OS crash.
I use the /var/lib/xen/images/box1 as the storage space of the guest OS. I
found that the file box1 is still there. So I was trying to use the
"restore saved machine" of virtual machine manager to reload it, However
it said that "Error restoring domain '/var/lib/xen/images/box1'. Is the
domain already running?"
SO I use the "xm list" to show all domains, only domain0 is there.
Anybody knows how to fix the above problem? Why did the newly created
virtual machine disappeared whenever I rebooted it? How to make it won't
disapper? Or how to load it into system again?
Regards,
Angel
17 years, 1 month
Problem about guest virtual machine creation
by taoj2@cs.rpi.edu
HI,
I wanna to do some experiments about VM migration. Currently I focus on
the local migration which means migrating a domain from one VM to another
VM in the same host.
First, I create two guest VMs, XEN1 and XEN2, on the same host. Both of
guests use FC6 as the operating system. And I also selected the XEN
package when I customize the FC6 installation.
Then, I was trying to create another VM, XENX, which is inside XEN1. The
layering relationship is like this:
HOST[VM1[VMX],VM2]
The host includes two VMs, VM1 and VM2. And I wanted to create another VMX
inside the VM1. Since the operating system of VM1 is FC6 and XEN is an
built-in application of VM1 I thought I could create VMX successfully.
However it prompted:
"Unable to open a connection to the Xen Hypervisor/daemon
Verify that:
-A Xen host kernel was booted
-The Xen service has been started"
Is there anything wrong with what I have done? Or currently we are not
allowed to create a virtual machine inside another VM. The only possible
way to create guest VM is to create it inside a physical machine, not VM?
Could you please let me know if you have any idea about above problem?
Thanks a lot.
Regards,
Angel
17 years, 1 month
Fedora 6 live migration
by Adrian Revill
Hi,
I have set up 2 machines called FedOne (4cpu)and FedTwo (2cpu) with FC6
and a third machine to act as an iSCSI target.
On each Dom0 i have started the iSCSI inititaor and ensured the devices
have identical names.
I have created a FC6 guest and can successfully start it on either
FedOne or FedTwo, so im confident the DomU is good.
I have enabled the migration interface and can start a migration from
either machine, but i have only been successful the very first time i
did it.
When i start the live migration from say FedOne to FedTwo, i see the new
DomU created and paused on FrdTwo, and the DomU enters the state Shutoff
(in virt-manager)
Then i can get 2 outcomes, it either just stays in that state, and the
DomU stays working.
Or the DomU migrates and then appears to hang (ie i cant ssh into it,
and existing ssh sessions hang) Also the console from virt-manager says
its unavailable, and the virt-manager on FedTwo shows the DomU using 50%
CPU,.
If i now migrate back to FedOne, it stays in a hung state, but also the
virt-manager reports 25% CPU usage.
Anyone got any ideas?
Also what should a normal migration do? When i ran it, it simply stopped
the DomU and created a new one on the other machine (booted it up)
Adrian
17 years, 1 month
Mount smb shares on domUs
by Asrai khn
Hi,
Few days back I have asked, how to boots domU in specific order so that a
domU which providing smb (samba) shares to other domu boots before any
other, I tried the approach of putting 'xm create xxx.cfg' in rc.local (not
using xendomines script) and I have put 'sleep 15' so that first one got
some extra time.
But apparently still few domUs sometime 1 or 2 boots before smb providing
domU and unable to mount the share. Actaully samba domU have to start more
services then other domU and this takes more time and i can't reduce these
services on boot coz they are needed.
I am wondering is there anything could be done at domUs side which are
mounting these smb shares so that they tries again if they can't find the
master from where they are mounting the shares.
may be something like ..
while (mount_samba()==unsuccessful) { sleep(5) }
But I am not sure where to stick it :)
Here is the line from /etc/fstab we are using to mount smb share..
//Master-H/samba/domU-b /var/mounthere cifs
username=xxxx-b,password=xxxxxxx 0 0
Any suggestion will be greatly appreciated.
Thanks. Askar
17 years, 1 month