[libvirt-users] Bypass USB to Xen VM
by Lorenzo Milesi
Hi. I'm trying to bypass an USB device from the host to a paravirtualized Xen VM.
I found this page:
http://libvirt.org/formatdomain.html#elementsUSB
which says how to bypass usb and pci devices to guests. It says that works on KVM only, but is that notice about PCI or about ALL?
Anyway I tried both from virt-manager and from virsh edit domain by adding the following:
<hostdev mode='subsystem' type='usb'>
<source>
<vendor id='0x03f0'/>
<product id='0x0114'/>
</source>
</hostdev>
but the edit is ignored, right after I save, if I do a dumpxml the "hostdev" stanza is not present.
Why? Any other way to bypass? usbip?
thanks
13 years, 2 months
[libvirt-users] Display port allocation
by Tom Hughes
The automatic port allocation algorithm for displays in libvirt seems to
assume that it is the only thing that will be allocating ports in the
normal VNC range starting at port 5900.
My host is running vino to export it's display, which has bound itself
to port 5900, yet libvirt is still trying to use port 5900 for both VNC
and Spice displays if I leave it in auto allocation mode.
I thought that this used to work in F14 but having just tried it again
it looks like I may have been mistaken.
Certainly it doesn't seem to work properly in F15 either with the
standard libvirt or with the virt-preview one.
Tom
--
Tom Hughes (tom(a)compton.nu)
http://compton.nu/
13 years, 2 months
Re: [libvirt-users] virsh slow on execution
by Lorenzo Milesi
> Also, when I start a VM via OCF, while guest is booting the command
> virsh domstate mydom
> returns "no state". This makes PaceMaker believe the resource is not
> running and continuously stops and restart the machine.
I've let a shell open with virsh domstate mydom active in loop, and it seems that it returns the correct state one time once very much requests! Say over 20 requests, only one or two says "idle" or "running", all the others return "no state".
I'd like to know why virsh fails so many times returning correct state...
thanks
13 years, 2 months
[libvirt-users] virsh slow on execution
by Lorenzo Milesi
Hi.
I've set up a Debian6+Xen4+libvirt0.9 server. VMs are going to be managed by PaceMaker with VirtualDomain primitive.
As found on some mailing list I've set
export VIRSH_DEFAULT_CONNECT_URI="xen:///"
in root's .bashrc, but when I run virsh commands, it takes up to 10s to get the result!
Is this normal?
At some point today I had issues connecting to libvirt, and in this case it was responding fast, less than couple of seconds.
Also, when I start a VM via OCF, while guest is booting the command
virsh domstate mydom
returns "no state". This makes PaceMaker believe the resource is not running and continuously stops and restart the machine.
How can I speed up virsh output?
Thanks
13 years, 2 months
[libvirt-users] Problems migrating xen domains
by Lorenzo Milesi
Hi.
I've set up a cluster with two Debian 6 nodes, running Xen 4 and libvirt 0.9 (from backports).
My VMs run over a DRBD resource, to allow live migration between the nodes.
I've successfully tried live migration using xm migrate, now I'd like to make it work using virsh, but cannot make it. Here's the error I'm getting:
# virsh migrate efw xen+ssh://10.22.22.152/
error: POST operation failed: xend_post: error from xen daemon: (xend.err "can't connect: (-2, 'Name or service not known')")
As said, xm migrate worked at first, so xend should be set up correctly.
I've done several searches but with no luck.
What could be the problem?
thanks
Relevant configs:
# cat /etc/xen/xend-config.sxp |grep -v ^#|grep -v ^$
(xend-http-server yes)
(xend-unix-server yes)
(xend-relocation-server yes)
(xend-port 8000)
(xend-relocation-port 8002)
(vif-script vif-bridge)
(dom0-min-mem 196)
(enable-dom0-ballooning yes)
(total_available_memory 0)
(dom0-cpus 0)
(vncpasswd '')
# env|grep VIRSH_DEFAULT_CONNECT_URI
VIRSH_DEFAULT_CONNECT_URI=xen:///
root@xensrv01:~# virsh version
Compiled against library: libvir 0.9.0
Using library: libvir 0.9.0
Using API: Xen 3.0.1
Running hypervisor: Xen 4.0.0
13 years, 2 months
[libvirt-users] basic network setup
by Jerry Worley
I am setting up a network with 2 guests, each on different subnets. The physical machine has 2 NICS, one on each subnet. I want the physical and guest machines to have IP addresses with no NAT, in their respective subnets, and then I want a separate subnet where the 2 guests can connect directly to each other. Can somebody explain conceptually the best way to use physical adapters, bridges, virtual bridges, virtual networks to accomplish this.
13 years, 2 months
[libvirt-users] Virsh Console - Changes in guest
by Abhishek Gupta
Hello
Virsh Console was not working on my setup. (also it required ctrl +
shift + ] instead of ctrl+alt+] for interrupt)
But after I create a file /etc/init/ttyS0.conf
# ttyS0 - getty
#
# This service maintains a getty on ttyS0 from the point the system is
# started until it is shut down again.
start on stopped rc RUNLEVEL=[2345]
stop on runlevel [!2345]
respawn
exec /sbin/getty -L 38400 ttyS0 vt102
Reload libvirt-bin, and restart the vm the console starts working.
Is there any other way of doing this without the need of changing the
guest files and independent of guest OS?
I am using libvirt version 0.8.8.
--
Abhishek
abhishekgupta92.info
13 years, 2 months
[libvirt-users] 2 virsh consoles on the same hypervisor causes conflicts
by David Lane
Anyone encountered this? Anyone have a solution? Seeing it in Fedora15
with KVM
If 2 consoles are opened to the same guest from the hypervisor they will
interfere with each other. keystrokes are lost and returned information is
displayed on either virsh console instances randomly. Both consoles are
rendered unusable.
Thanks,
DAVID
13 years, 2 months
[libvirt-users] Virtual switches support (Openvswitch, VDE,...)
by Jérémie Tarot
Hi,
IIR/UC there have been a couple of discussions lately about possible
evolution of libvirt's network management features to eventually support
more kinds of virtual networking technologies like Openvswitch, VDE,...
Could a kind and informed soul keep us on touch with progress in the field ?
Thank you very much in advance
PS: I wish this will be available before/for 1.0 release...
13 years, 2 months
Re: [libvirt-users] [Qemu-devel] Qcow2
by Stefan Hajnoczi
On Wed, May 18, 2011 at 5:18 PM, <arun.sasi1(a)wipro.com> wrote:
> Hello Stefan,
>
>
>
> Thank you very much for considering my issue...
>
>
>
> Here is My problem...
>
>
>
> 1) I have 4 VM`s Running on Base server.
>
> 2) Base server is having 15GB RAM.
>
> 3) I can start all VM`s apart from my file server.
>
> 4) File server is having OS image and 1.2TB QCOW2 Image.
>
> 5) When I remove 1.2TB from xml file (vdb.img) I can start my server but
> Data partition is null becase I have not added 1.2TB in XML.
>
> 6) Command I am using virsh start <Domain>
>
> 7) Even through virt-manager also I am not able to start the VM. Attached is
> the error message
Unfortunately the message gives no details:
"libvirtError: internal error unable to start guest:"
> 8) Keep on trying adding and removing and defining the XML after 2 Hrs I am
> able to start.
What did you change that made it work?
> 9) Operating system Ubuntu 9.04
>
> 10) I have 3 location file server having the same issue… even all file
> servers are same configuration…
Libvirt keeps log files for each VM under
/var/log/libvirt/qemu/<domain>.log. You might find more detailed
error information in that file and it will at least contain the
qemu-kvm command-line used to launch the domain. Please take a look
at that file and post the lines surrounding the error, especially the
"starting up" log line.
Have you verified that the .qcow2 image file is accessible, ownership
and permissions should be the same as the image files of your other
domains.
Stefan
> -----Original Message-----
> From: Stefan Hajnoczi [mailto:stefanha@linux.vnet.ibm.com]
> Sent: Wednesday, May 18, 2011 9:28 PM
> To: Arun Sasi V (WI01 - Manage IT)
> Subject: Re: Qcow2
>
>
>
> On Wed, May 18, 2011 at 07:30:05PM +0530, arun.sasi1(a)wipro.com wrote:
>
>> The Guest VM which is larger example 1.2TB is not starting after the
>
>> base server reboot... Pleae help what needs to be do in this situation.
>
>
>
> You have not provided information on what is happening. Are you getting
>
> an error message during boot from the 1.2TB guest? Is the 1.2TB guest
>
> even being autostarted (use "virsh autostart <domain>" to enable
>
> autostart on a domain)?
>
>
>
> Please send support questions like this to qemu-devel(a)nongnu.org in the
>
> future. That way other people in the QEMU community can help too.
>
>
>
> Stefan
>
> Please do not print this email unless it is absolutely necessary.
>
> The information contained in this electronic message and any attachments to
> this message are intended for the exclusive use of the addressee(s) and may
> contain proprietary, confidential or privileged information. If you are not
> the intended recipient, you should not disseminate, distribute or copy this
> e-mail. Please notify the sender immediately and destroy all copies of this
> message and any attachments.
>
> WARNING: Computer viruses can be transmitted via email. The recipient should
> check this email and any attachments for the presence of viruses. The
> company accepts no liability for any damage caused by any virus transmitted
> by this email.
>
> www.wipro.com
13 years, 2 months