[libvirt-users] virt-manager/libvirt backwards compatibility problem?
by Whit Blauvelt
Hi,
Wanting to test some recent features of libvirt, I installed 0.9.3 on a
couple of systems. It works fine on those, but since upgrading neither of
them can successfully connect with virt-manager or virsh to a couple of
other systems running libvirt 0.8.3. Even after upgrading virt-manager and
virtinst to the latest versions on the 0.9.3 systems, they fail like this
in virt-manager:
Unable to open a connection to the libvirt management daemon.
Libvirt URI is: qemu+ssh://root@192.168.1.67/system
Verify that:
- The 'libvirtd' daemon has been started
Cannot recv data: : Connection reset by peer
Traceback (most recent call last):
File "/usr/src/virt-manager-0.8.7/src/virtManager/connection.py", line
1055, in _try_open
None], flags)
File "/usr/local/lib/python2.7/dist-packages/libvirt.py", line 102, in
openAuth
if ret is None:raise libvirtError('virConnectOpenAuth() failed')
libvirtError: Cannot recv data: : Connection reset by peer
Virsh simple fails with:
error: Cannot recv data: : Connection reset by peer
error: failed to connect to the hypervisor
Since it's certain that the libvirtd daemon has been started on the other
systems (they have production VMs that are running), this leaves me puzzled.
This connection for virt-manager worked flawlessly for months when all the
systems had libvirt 0.8.3.
Is this a known incompatibility? Something I have to tweak somewhere?
Thanks,
Whit
13 years, 5 months
[libvirt-users] Compatibility with Parallels Virtouzo
by Andreas Mauf
Hi,
will libvirt work with the commercial parallels virtuozzo, too? Cause
the linux version of virtuozzo is based on openvz, which ist supported
by libvirt.
Does someone have some experencies with that?
Thanks for any feedback,
Andreas
--
SysEleven GmbH
Rosenthaler Str. 34/35
10178 Berlin
Firmensitz: Berlin
Registergericht: AG Berlin Charlottenburg, HRB 108571 B
Geschäftsführer: Marc Korthaus
13 years, 5 months
[libvirt-users] Questions on libvirt hook scripts
by Whit Blauvelt
Hi,
Reading http://libvirt.org/hooks.html I'm wondering of the following
implication is accurate:
Under "QEMU guest migration" it says
At the beginning of the migration, the qemu hook script on the
_destination_ host is executed with the "start" operation.
Above that, under "Specifics"
Before a QEMU guest is started, the qemu hook script is called in
two locations; if either location fails, the guest is not started.
The first location, since 0.9.0, is before libvirt performs any
resource labeling.... This is called as:
/etc/libvirt/hooks/qemu guest_name prepare begin -
The second location, available Since 0.8.0, occurs after libvirt has
finished labeling all resources, but has not yet started the guest,
called as:
/etc/libvirt/hooks/qemu guest_name start begin -
Is it accurate that under 0.9.x, starting a VM will run the qemu hook script
as both "prepare" and "start," but migrating a VM will only run the qemu
hook script as "start"? Or is the description incomplete with libvirt also
running "prepare" on migration?
Another question, under "Calling libvirt functions from within a hook
script" it says
DO NOT DO THIS!
A hook script must not call back into libvirt, as the libvirt daemon
is already waiting for the script to exit.
A deadlock is likely to occur.
That makes obvious sense. But does it apply to calling libvirt on a second
system? For example, a script might check the status of VMs on a second
system as a condition of starting a VM on first, using virsh. My guess is
that shouldn't deadlock. But then since during migration "the qemu hook
script on the _destination_ host is executed," if the destination hook
script were in turn to use virsh to check the _source_ host, then in a
roundabout way a libvirt function has been called from within the hook
script. I can't quite picture whether this particular instance would be
dangerous.
If it's accurate that the _destination_ hook script is only called with
"start" and not with "prepare," the second question matters less. The checks
on the second system that I want to do can be limited to the "prepare"
invocation, and so never loop back.
On the other hand, if the destination hook script _is_ called with "prepare"
during migration, it would be useful to know how to tell if it's being
called as part of a migration.
TIA,
Whit
13 years, 5 months
[libvirt-users] NPIV + KVM
by Eric Ross
Is this possible:
I would like to have a VM use a FC LUN exported to a virtual wwn (using
NPIV) as its primary disk.
I was thinking there might be a mechanism to say VM, boot using a LUN
with this vwwn as your primary disk?
Thanks in advance.
--
---------------------------------------------------------------------
Eric Ross
HPC Systems Administrator
Argonne National Laboratory
Leadership Computing Facility
---------------------------------------------------------------------
13 years, 5 months
[libvirt-users] Creating a CENTOS VM
by Dale Amon
I used the following command:
virt-install --connect qemu:///system -n server -r 512 --vcpus=1 -f kvm/server.cyberteams.com/root.qcow2 -s 12 -c iso/CentOS-5.6-x86_64-bin-DVD-1of2.iso --vnc --noautoconsole --os-type linux --accelerate --network=bridge:br0 --hvm
which with a BlueOnyx CD some months ago worked fine. However
when I connected with virt-viewer, after going through the TEST
function it then claimed it could not find the CD.
Is CENTOS a bit strange in some way?
13 years, 5 months
[libvirt-users] Changing path of VM configuration files
by Jatin Kumar
Hello,
I am using shared storage and i want to keep all the configuration files of
vms of different hosts in one single location (on shared storage).
Is there a way i can change the default path=/etc/libvirt/qemu/<VMName>.xml
to something else.
I looked into libvirtd.conf and qemu.conf but could not find any such thing.
--
Jatin
13 years, 5 months
[libvirt-users] KVM and <interface type='ethernet'>
by Maximilian Wilhelm
Hi!
I'm trying to set up a VM with "just an tap device" with the following config
(basicly inherited from 'virsh domxml-from-native qemu-argv'):
<domain type='kvm'>
[...]
<interface type='ethernet'>
<mac address='52:54:1f:ac:08:51'/>
<target dev='foo_eth0'/>
<model type='virtio'/>
</interface>
[...]
</devices>
</domain>
This sadly failes with:
| error: Failed to start domain foo
| error: internal error process exited while connecting to monitor: could not configure /dev/net/tun: Device or resource busy
I'm wondering why this fails as there shouldn't anything be done with
the tap device besides setting it up.
(The basic goal would be to set up a VM with a tap device, put an v4
and v6 IP on it and let OSPF do the rest. As the <script> parameter
will not work with dropped privileges I wrote a wrapper for the kvm
binary which takes care of this via sudo. This part works, but I fail
to get the VM started with <interface type='ethernet'>)
Any ideas?
Thanks
Max
--
<@uschebit> Das erklärt die Sitte. Das ist so ähnlich wie mit dem Verbeugen zur Begrüßung in Japan.
Das dient auch nur dazu, durch Begutachten der Schuhqualität den sozialen Rang festzustellen.
13 years, 5 months
[libvirt-users] SSH console for qemu
by Nikita A Menkovich
Hello,
I'm using qemu with libvirt and I want to create SSH emergency console
to linux/unix guests, same way as it is realized in qemu VNC.
Maybe someone could help me to determine a place to dive in this question.
I think the best way to make it - extend qemu API. But if this could
be made with libvirt - this will be also fine.
--
Nikita A Menkovich
http://libc6.org/
JID: menkovich(a)gmail.com
13 years, 5 months
