[libvirt-users] sluggish behavior
by Ainsworth, Thomas
Greetings,
This is my first stab at this mailing list. Forgive me if I am over/under
verbose.
We have been running KVM for a while now and love it. Recently I have
noticed a weird anomaly
that is getting on my nerves. When I clone (GUI or virt-clone from CLI)
both* virt-manager* and *virsh*
enquiries get extremely sluggish. Using top I cannot see any resource
issues. No I/O wait, the load
average show noting outside of the normal, nothing taking CPU or Memory. I
have notice that in a
couple of the instances the libvirtd daemon was stopped. We have more
than twenty (20 +) VM's
running - they all seem fine. The metal server is a beast of an 820 with
192GB of RAM and 64 CPU's.
The load average rarely creeps above 2 or 3. Do not think it is a resource
issue. I can write a test
file to the filesystem/volume where the images live at over 350MB/s.
The sluggishness disappears after a reboot. If I clone a VM ... back again
...
Any ideas?
Thanks in advance,
Tom Ainsworth
RHCSA
10 years, 5 months
[libvirt-users] [LXC] feature request - hostdev target
by Piotr Bartosiewicz
Hi,
I am trying to mount /dev/tty5 from host as /dev/tty1 inside LXC container,
but libvirt currently supports only 1-to-1 mapping:
<hostdev mode='capabilities' type='misc'>
<source>
<char>/dev/tty5</char>
</source>
</hostdev>
What do you thing about extending this by adding target parameter?
It is easy to archive manually using mknod (mknod /dev/tty1 c 4 5),
but still I think it would be nice to add this feature to the libvirt.
Regards,
Piotrek
--
Piotr Bartosiewicz
Samsung R&D Institute Poland
Samsung Electronics
10 years, 5 months
[libvirt-users] Need help with [virt net-create]
by Srinivas_G_Gowda＠dell.com
Is it possible to configure 2 virtual nics in the same ip range ?
Here is what I am trying.
**** This works ****************
cat nw1.xml
<network>
<name>nw1</name>
<domain name='nw1' />
<forward mode='nat' dev='eth2'/>
<ip address='192.168.101.1' netmask='255.255.255.0'>
<dhcp>
<range start='192.168.101.128' end='192.168.101.254'/>
</dhcp>
</ip>
</network>
[root@localhost nw]# virsh net-create nw1.xml
Network nw1 created from nw1.xml
[root@localhost nw]# virsh net-list --all
Name State Autostart Persistent
----------------------------------------------------------
nw1 active no no
default active yes yes
****** FAILS when given the same ip address range of an existing virtual interface *******
[root@localhost nw]# cat nw2.xml
<network>
<name>nw2</name>
<domain name='nw2' />
<forward mode='nat' dev='eth2'/>
<ip address='192.168.101.1' netmask='255.255.255.0'>
<dhcp>
<range start='192.168.101.128' end='192.168.101.254'/>
</dhcp>
</ip>
</network>
[root@localhost nw]# virsh net-create nw2.xml
error: Failed to create network from nw2.xml
error: internal error: Network is already in use by interface virbr1
****** PASS - on giving diff ip range *******
[root@localhost nw]# cat nw3.xml
<network>
<name>nw3</name>
<domain name='nw3' />
<forward mode='nat' dev='eth2'/>
<ip address='192.168.102.1' netmask='255.255.255.0'>
<dhcp>
<range start='192.168.102.128' end='192.168.102.254'/>
</dhcp>
</ip>
</network>
[root@localhost nw]# virsh net-create nw3.xml
Network nw3 created from nw3.xml
[root@localhost nw]# virsh net-list --all
Name State Autostart Persistent
----------------------------------------------------------
nw1 active no no
nw3 active no no
default active yes yes
Any suggestions on what could be wrong here ?
Thanks,
G
10 years, 5 months
[libvirt-users] Cloning of Running Domain
by Thomas
Hello,
I have a running transient domain with a storage volume that utilizes a
backing file (both QCOW2 format).
I'd like to capture the disk and memory state of the domain preferably to
external file(s) (which I understand is an 'external checkpoint'), so that
other independent domains can be instantiated and effectively "resumed"
from this point-in-time profile.
My network configuration is such that domains are network-isolated at L2,
so resolving MAC/IP address and similar conflicts within the guest are not
of concern.
Is it possible to accomplish this? I've looked at both the
snapshot/checkpoint and migration APIs and it appears as though the pieces
are there, but am having difficulty defining the proper workflow.
For instance, an external online checkpoint can be created using
virDomainSnapshotCreateXML, but is it possible to disassociate the snapshot
from the source domain (effectively returning it to its pre-snapshot state)
and at a later time instantiate new domains using the the disk/memory data
from the snapshot files?
Any recommendations, insight, or guidance is appreciated.
Thanks,
-Thomas
10 years, 5 months
[libvirt-users] routed mode
by JvR O
Hi,
I have installed a centos65_guest_1 system on a centos65 host with
virt-manager.
By default libvirt has a default network in NAT mode. At centos65_guest_1
when pinging to google it works.
But I would like to use a routed mode for production enviroment with some
services online,
like http, ftp, ssh, etc.
I have created the virtual network 192.168.100.0/24 with routed mode, and I
have installed another guest system, centos65_guest_2, that is using this
network.
The problem is that I can't ping to the outside world.
ping to 192.168.100.1 is working
Why is not working?
Any suggestion?
Thank you
10 years, 5 months
[libvirt-users] help to implement LXC terminal
by Fl＠sh
Hi, guys!
I began to write qt-virt-manager
(https://github.com/F1ash/qt-virt-manager), using libvirt API.
I'm currently trying to implement a terminal for
LXC-containers, but I do not have enough knowledge and
experience. As I understand it, i must use
virDomainOpenConsole, which binds virStream with working
domain (Virtual Machine). But how to use this stream, I do
not know ( I found qterminalwidget to implement the terminal,
but I do not know how to tie them). Can anyone suggest how
can I do?
--
Fl@sh
10 years, 5 months
[libvirt-users] Invoking virsh console from Java
by Nagaraj Mandya
Hello,
I have a Java application from which I am invoking the "virsh console"
command to access the console of a VM. I invoke the virsh command using
ProcessBuilder.start(). However, I am unable to communicate with the stdin
of the VMs console through the OutputStream of the Process object.
When I invoke "virsh console" from within Java, I see the following
messages on stderr:
info : libvirt version: 0.9.8
error : vshRunConsole:318 : unable to get tty attributes: Inappropriate
ioctl for device
Does anybody know how I can workaround this? Thanks.
--
Regards,
Nagaraj
10 years, 5 months
[libvirt-users] convert physical windows 8 machine to virtual machine
by arnaud gaboury
I dual boot w8 and Arch Linux. Both are on the same ssd drive (each OS
have of course a partition).
Now I would like to virtualize w8 and run it inside Arch Linux using
KVM/QEMU and Libvirt.
w8 has already been installed on a ntfs partition. As it is brand new,
it will not be difficult to reinstall it on a image.raw
I have been reading some articles and found myself a little bit confused.
1- Some articles point to driver issues on the windows size, which
could lead to an unusable OS. Is this a real issue ? Workaround found
is to uninstall IDE/SATA driver in w8 and then install
general/universal driver.
2- I can of course copy my w8 partition to a image.raw in my second
HD, then create the VM. But in this case, after time, I will not have
the same w8 on my ssd and my VM. What I am looking for is a VM that
will ever replicate w8 on ssd. In this case, I guess I need to
reinstall w8 on a raw qemu-img then use this image for the VM. Am I
correct ? Is there anyother solution ?
Is it worth keeping the raw image on the ssd for better performance,or
moving the image to the other HD will make no difference?
3- Most of the reading talk about virtualise a physical windows using
Virtualbox or VMware, which I certainly don't want. I would appreciate
any good pointer for the overall set-up.
Thank you in advance for hints, tips and advices.
10 years, 5 months
[libvirt-users] aliases for graphic devices
by Francesco Romani
Hello,
AFAIK, aliases cannot be set for Graphics devices. I tried this very simple experiment
with libvirt 1.2.4 (from virt-preview on F20)
relevant part of the input xml:
<graphics type='spice' port='-1' autoport='yes' listen='127.0.0.1'>
<listen type='address' address='127.0.0.1'/>
<alias name='display.0'/>
</graphics>
<graphics type='vnc' port='-1' autoport='yes' listen='127.0.0.1'>
<listen type='address' address='127.0.0.1'/>
<alias name='display.1'/>
</graphics>
# virsh -c qemu:///system
virsh # create testdom_spice_vnc_aliases.xml # see attachment
Domain fedora19 created from testdom_spice_vnc_aliases.xml
virsh # dumpxml fedora19 # only relevant part, see attachment for full output
[...]
<graphics type='spice' port='5900' autoport='yes' listen='127.0.0.1'>
<listen type='address' address='127.0.0.1'/>
</graphics>
<graphics type='vnc' port='5901' autoport='yes' listen='127.0.0.1'>
<listen type='address' address='127.0.0.1'/>
</graphics>
[...]
So, are really aliases not allowed for the Graphics Devices or am I doing something wrong?
If aliases aren't allowed, any plans to support them?
Thanks and best regards,
--
Francesco Romani
RedHat Engineering Virtualization R & D
Phone: 8261328
IRC: fromani
10 years, 5 months
[libvirt-users] Is there a way to get host uptime in remote libvirt
by coffeeball
We manage hypervisors (VMWare ESXi/vCenter, KVM, XEN, Hyper-V) by remote Libvirt API, in our case we need to get the host uptime via the same libvirt interface. Is there a way get this info now for all the aforementioned hypervisor types?
The APIs virConnectGetSysinfo(), virNodeGetInfo() provide host info but it doesn't include the system uptime.
The virNodeGetCPUStats() can returns CPU usage in nanosecond, can we add the user + system + idle + iowait to calculate the system uptime? Looks like the sum value has a huge gap with the real uptime value returned by "uptime" CLI.
virsh # nodecpustats
user: 2985046540000000
system: 1368209710000000
idle: 388018933920000000
iowait: 43371960000000
10 years, 5 months
