[libvirt-users] Question about virtio drivers
by pa1n_0
Hello,
Any ideas from which version of virtio drivers support well high-end
configuration like >= 8 physicals CPUs and >=24Go.
I know that using the last version is the way to do but someone would like
to know an accurate version.
Thanks for answer.
10 years, 9 months
[libvirt-users] "connection refused", but server is listening
by Pasquale Dir
netstat -lptu
gives me
tcp6 0 0 [::]:16514 [::]:* LISTEN
1314/libvirtd
so the server is correctly listening on interfaces.
My /etc/libvirt/libvirtd settings are defaulted.
My /etc/default/libvirt-bin has "-d -l" options so server is listening.
If I do virsh -c qemu://143.225.229.190/system (that is my ip or an ip
belonging to my network) I get "connection refused".
How come?
And note that if I put my address, 143.225.229.190, as listening address in
the libvirtd conf file, I get "this address class is not supported".
10 years, 9 months
[libvirt-users] create ovs port without root
by Vasiliy Tolstov
Hello! How can i operate with openvswitch without root rights?
For example - i can add my user to kvm group and create vm from
libvirt with my own user, but now i'm failed with creating port.
errors in logs:
Mar 06 14:04:46 selfip.ru ovs-vsctl[19065]:
ovs|00001|vsctl|INFO|Called as ovs-vsctl --timeout=5 -- --if-exists
del-port
Mar 06 14:04:46 selfip.ru ovs-vsctl[19065]:
ovs|00002|vsctl|ERR|'del-port' command requires at least 1 arguments
Mar 06 14:04:46 selfip.ru libvirtd[6418]: internal error: Child
process (ovs-vsctl --timeout=5 -- --if-exists del-port) unexpected
exit status 1: ovs-vsctl: 'del-port' command requires at least 1
arguments
Mar 06 14:04:46 selfip.ru libvirtd[6418]: Unable to delete port (null)
from OVS: Operation not permitted
--
Vasiliy Tolstov,
e-mail: v.tolstov(a)selfip.ru
jabber: vase(a)selfip.ru
10 years, 10 months
[libvirt-users] ISO refuses to let vm starts (and is not mentioned in config)
by Mauricio Tavares
Trying to start one of my vms, a centos one at that, but am
getting the following message:
[root@vmhost ~]# virsh start voip --paused
error: Failed to start domain voip
error: cannot open file
'/var/tmp/FreePBX-5.211.65-3-x86_64-Full-1388073872.iso': No such file
or directory
[root@vmhost ~]#
But, virsh dumpxml voip shows no info onto the .iso
<disk type='file' device='disk'>
<driver name='qemu' type='raw' cache='none' io='native'/>
<source file='/dev/vmhost_vg0/voip'/>
<target dev='vda' bus='virtio'/>
<address type='pci' domain='0x0000' bus='0x00' slot='0x04'
function='0x0'/>
</disk>
<disk type='file' device='cdrom'>
<driver name='qemu' type='raw'/>
<target dev='hda' bus='ide'/>
<readonly/>
<address type='drive' controller='0' bus='0' target='0' unit='0'/>
</disk>
<disk type='file' device='cdrom'>
<driver name='qemu' type='raw'/>
<target dev='hdb' bus='ide'/>
<readonly/>
<address type='drive' controller='0' bus='0' target='0' unit='1'/>
</disk>
Log file is not particularly helpful:
2014-03-04 22:57:46.121+0000: shutting down
2014-03-05 23:42:18.528+0000: shutting down
2014-03-05 23:47:07.375+0000: shutting down
2014-03-06 00:56:32.908+0000: shutting down
2014-03-06 01:11:32.582+0000: shutting down
So, the question is how did this iso got involved in this specific vm?
I can start my other vms without them even knowing about this .iso.
10 years, 10 months
[libvirt-users] Attaching a hostdev on startup
by Mat Arge
Hello!
I have written a script which attaches some USB tokens as hostdev devices
to (qemu) VMs managed by libvirt. This works fine as long as I call the
script manually.
But what I really want to do, is run this scrip from a startup hook, so
that whenever a Vm is started, it looks if there are appopriate tokens
available for attachment. But when the script is run from the hook, it
(along with the whole libvirtd) freezes. That doesn't really come as a
surprise, since http://libvirt.org/hooks.html explicitely states
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.
How can I get around this problem?
cheers
Mat
10 years, 10 months
[libvirt-users] Providing fallback DNS servers via DHCP
by Florian Rüchel
Hello everyone,
TL;DR: Is ist possible to provide a secondary DNS (e.g. 8.8.8.8) via
DHCP in addition to the gateway (in my case 10.0.1.1)?
I use libvirt to manage all my qemu-based VMs. I have set up a local
NAT-network which includes all my VMs (and some physical servers
integrated via VPN). Everything works fine, but ocassionally, I have to
restart the network (when I add instances for example). In that case, my
VMs loose their local network connection until I restart them.
However, some instances can not be restarted at will and so a change in
network configuration means they loose their local connectivity.
However, they still have a second network device that has an external
static IP. This IP can still connect and be connected to, so it is
feasible to loose the connection on the local network.
Unfortunately though, libvirt (through dnsmasq's DHCP) broadcasts itself
as the only DNS. This results in name resolution going down. If I add
e.g. 8.8.8.8 to resolv.conf, I can resolve names again (except local
ones, of course). So it would be nice if the DHCP would directly supply
additional DNS servers.
For dnsmasq, I would add to the configuration something like
dhcp-option=6,0.0.0.0,8.8.8.8
However, looking through the archive, I found that you guys prefer to
keep it implementation agnostic and thus you do not provide the option
to set this directly (since this is dnsmasq-only).
So I was wondering if there is any way to achieve this with libvirt? Is
there already an implemented way I could not find? Or is there maybe a
*generic* option to provide some kind of extra-configuration specific to
the implementation? Or is my only solution to make sure the clients
manually add a second DNS by themselves?
Regards,
Florian
10 years, 10 months
[libvirt-users] Migration failure after libvirt upgradation
by Shweta Salaria
Team,
I'm trying to migrate a domain offline(shutdown mode) using libvirt-python
API 'migrateToURI' but it gives me an error :
libvirt: QEMU Driver error : Requested operation is not valid: domain is
not running
When I tried the same API for live migration(running and suspend mode), it
gives me an error:
libvirt: QEMU Driver error : Unsafe migration: Migration may lead to data
corruption if disks use cache != none
Below are the versions of the virtualization softwares I am using currently
:
Hypervisor : KVM
qemu : 1.0
qemu-kvm : 1.0
Libvirt : 1.2.1
Earlier, I had performed offline and live migration *successfully* with
these versions:
Hypervisor : KVM
qemu : 1.0
qemu-kvm : 1.0
libvirt : 0.9.8
Also, I would like to know whether it's possible to migrate a domain having
snapshots?
Kindly guide me through this.
Thanks in advance
Regards,
Shweta
10 years, 10 months
[libvirt-users] How to enable mount loop device in LXC using libvirt LXC driver
by Du Jun
When I try to mount loop device in a linux container, I got the following
error message:
mount: Could not find any loop device. Maybe this kernel does not know
about the loop device? (If so, recompile or `modprobe loop'.)
I wonder how to enable mount loop device in LXC using libvirt LXC driver?
--
Thanks!
Du Jun
10 years, 10 months
