- Users - libvirt List Archives

[libvirt-users] guest VM waiting for a symbol at serial console
by Andrey Korolyov 09 Jul '12

09 Jul '12

Hi, Recently I`ve run into weird bug, which may be reproduced in following way: - start qemu virtual machine configured with direct boot and kernel args like 'console=tty0 console=ttyS0,38400n8', connect to serial console, disconnect and then shutdown vm - try to start vm again, machine hangs on booting kernel until I press any key into 'virsh console' session - after first power cycle, this bug remains until host system will be rebooted. This happens only when vm experiencing 'power cycle', rebooting inside same kvm process do fine. guest kernel 3.4.4/3.2.20, libvirt 0.9.12 and qemu-1.1.0

1 0

[libvirt-users] numa query
by kapil jain 09 Jul '12

09 Jul '12

Hi, I have a host machine which contains 4 numa nodes each with 2 GB memory and 4 cpus. I am using qem-kvm hypervisor. I am trying to create a guest with the similar topology (4 numa nodes, with only 2 vcpus and 1 GB memory) as host. Each vcpu is 1-1 pinned to physical cpus(i.e. guest{socket0 cpu0} is pinned to host{socket0 cpu0). But one critical requirement is that one guest socket should not have memory from two host sockets to avoid numa access completely. With the current constructs of numa/topology I am able to create the guest. But memory is not mapped as socket 1-1. numatune is not helping. Please suggest a possible way. Thanks, Kapil

2 1

[libvirt-users] Metadata accessible within guests
by Renich Bon Ciric 09 Jul '12

09 Jul '12

Hello, I'd like to know if there is any way to make metadata; for example, the public part of an SSH, accessible within a Guest? This would be very useful in order to provide access to Guests in an automated way. For example, let's suppose I have an implementation that gives a user the power to create Guests. The user has stored his public SSH key in my implementation. After the user creates a Guest, I can inject (with a proper rc.local script or something of the sorts) his/her SSH key properly. This would be useful for a million things. Another cool feature would be that I could sync the root password with the SPICE/VNC password. anyway, if there is a way to store metadata into an XML and, maybe, access it with some tool from within the Guest, this would be awesome. Thank you for your time. -- It's hard to be free... but I love to struggle. Love isn't asked for; it's just given. Respect isn't asked for; it's earned! Renich Bon Ciric http://www.woralelandia.com/ http://www.introbella.com/

1 0

[libvirt-users] Why the xen vm memory won't bloon
by Tony Huang 09 Jul '12

09 Jul '12

Hello, I have a vm with startup xml file has currentmemory = 256M and memory=1024M. However, once the vm started up, the memory allocation won't change. xentop shows that the memory keeps in 256M. How can I make the vm to allocate more memory when they needed. -- Tony

1 1

[libvirt-users] CPU usage by libvirtd
by Zhihua Che 06 Jul '12

06 Jul '12

Hi, I employed libvirt to monitor 50 domains in one node(host). my monitor calls libvirtd for every stats such as cpu, memory, disk every 10 sec. All my requests are made in one thread. And I found that the libvirtd may cost 30% or above cpu time. Is it normal in my case? I tried to reduce this cpu consumption and found some variable in libvirtd.conf and got a little confused by variables like max_clients, max_workers, max_requests. In my opinion, max_clients limits the number of virConnect created by app. max_workers limits the number of thread to handle requests from clients. Does that mean requests from one client may be handled in multi-thread way if max_workers is greater than max_workers? max_request limits the concurrent RPC calls according the comments in libvirtd.conf. I wonder if it can improve the response time if I give a higher value in the case that all my requests are made in one thread? I've tried several different configurations and they didn't reduce the cpu usage effectively.

1 0

[libvirt-users] Problem halting/restaring a lxc container from within
by david＠dit.upm.es 05 Jul '12

05 Jul '12

Hi, I've been making some tests with libvirt and LXC and found some problems when halting/restarting a LXC container from within. Basically, on a Ubuntu 12.04 system with libvirt installed as package (0.9.8), I've created a basic container image with: lxc-create -t ubuntu -n lxc And started it using the libvirt XML listed below and the following command: virsh -c lxc:// create lxc.xml I can access the container console normally but when I issue a halt or reboot command from inside, the container initiates the halt/reboot execution but it does not finishes it properly. For example: $ sudo halt -p [sudo] password for ubuntu: Broadcast message from ubuntu@my-container (/dev/pts/0) at 23:19 ... The system is going down for power off NOW! ubuntu@my-container:~$ acpid: exiting * Asking all remaining processes to terminate... [ OK ] * All processes ended within 1 seconds.... [ OK ] * Deconfiguring network interfaces... [ OK ] * Deactivating swap... [fail] * Unmounting weak filesystems... [ OK ] umount: /run/lock: not mounted mount: / is busy * Will now halt The container seems to finish the shutdown process but libvirt does not seem to be signaled about it (virsh shows the container is still executing). Something similar happens with reboot. However, if a start that container with: lxc-start -n lxc and do the same test, it works perfectly: # sudo halt -p [sudo] password for ubuntu: Broadcast message from ubuntu@my-container (/dev/lxc/console) at 23:17 ... The system is going down for power off NOW! ubuntu@my-container:~$ acpid: exiting * Asking all remaining processes to terminate... [ OK ] * All processes ended within 1 seconds.... [ OK ] * Deconfiguring network interfaces... [ OK ] * Deactivating swap... [fail] umount: /run/lock: not mounted mount: cannot mount block device /dev/disk/by-uuid/9b50a43d-98c3-45ad-a540-7fcbc629a418 read-only * Will now halt # Any idea about how to investigate/solve this problem?? Thanks in advance, David Fernandez -- lxc.xml file -- <domain type='lxc'> <name>lxc</name> <memory>524288</memory> <os> <type>exe</type> <init>/sbin/init</init> </os> <vcpu>1</vcpu> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/lib/libvirt/libvirt_lxc</emulator> <filesystem type='mount'> <source dir='/var/lib/lxc/lxc/rootfs'/> <target dir='/'/> </filesystem> <interface type='network'> <source network='default'/> </interface> <console type='pty'/> </devices> </domain>

2 1

[libvirt-users] Reboot KVM guest failed via virsh!
by GaoYi 03 Jul '12

03 Jul '12

Hi, I have tried to reboot a KVM guest by this command line: virsh reboot vm1 However, the virsh report: reboot is not supported without json monitor. I am using libvirt-0.9.9 and qemu-0.14.0. Can anyone provide some help? Thanks, Yi

2 1

[libvirt-users] using vmchannel between 6.x host/guests
by Michael MacDonald 03 Jul '12

03 Jul '12

Hi all. Having trouble figuring out the magic to set up VMChannel comms between EL6.x host/guests. My end goal is to enable fence_virt in the guest to talk to fence_virtd on the host via VMChannel. I'd prefer to use that instead of multicast because it is supposed to work even if networking in the guest is down/borked. My analysis is that there is a mismatch between what libvirt is feeding qemu-kvm and what qemu-kvm is willing to accept: virt-install option: --channel unix,path=/var/run/cluster/fence/foobar,mode=bind,target_type=guestfwd,target_address=10.0.2.179:1229 turns into this XML: <channel type='unix'> <source mode='bind' path='/var/run/cluster/fence/foobar'/> <target type='guestfwd' address='10.0.2.179' port='1229'/> </channel> Which then gets fed to qemu-kvm as this: -chardev socket,id=charchannel0,path=/var/run/cluster/fence/foobar,server,nowait -netdev user,guestfwd=tcp:10.0.2.179:1229,chardev=charchannel0, id=user-channel0 And then qemu-kvm barfs like so: qemu-kvm: -netdev user,guestfwd=tcp:10.0.2.179:1229,chardev=charchannel0,id=user-channel0: Invalid parameter 'chardev' Versions: libvirt-0.9.10-21.el6.1 qemu-kvm-0.12.1.2-2.295.el6 NB: I did try this with the versions in 6.2, got the same result. Upgraded to these versions to see if the problem went away, but no joy. NB2: The fence_virt.conf manpage lists the following as example XML for defining a channel device: <channel type=’unix’> <source mode=’bind’ path=’/sandbox/guests/fence_molly_vmchannel’/> <target type=’guestfwd’ address=’10.0.2.179’ port=’1229’/> </serial> Any advice would be greatly appreciated. I can fall back to multicast, of course, but I'd like to make this work if possible. Thanks! Mike

1 0

[libvirt-users] idle connections (?)
by Evaggelos Balaskas 02 Jul '12

02 Jul '12

Hello, i have noticed that for some reason my windows 2003 & windows 2008 VMs drops map drives caused by idle connections. I am not a windows admin and i have no knowledge of troubleshooting this, except reading support.microsoft.com The win admin is yelling that this is a fedora17 libvirt problem and that i need to fix this. I have two different net cards and i am using macvtap with vepa model for the VMs Is there any possibility that could be true ? Can i test somehow or prove that this is not related with libvirt ? Thanks in advance for any help. -- Evaggelos Balaskas - Unix System Engineer http://gr.linkedin.com/in/evaggelosbalaskas

1 0

[libvirt-users] virsh iface-list () - function is not supported
by Ananth 02 Jul '12

02 Jul '12

Hi, I am using KVM hypervisor on Ubuntu 12.04 (libvirt-0.9.8). The call to list the interfaces seem to be failing on this version of libvirt as seen below. virsh # iface-list error: Failed to list active interfaces error: this function is not supported by the connection driver: virConnectNumOfInterfaces I have seen this issue with libvirt 0.9.2 as well, which seemed to be working in 0.9.4 version. Is there any work around for this or any alternate way of querying the interfaces and bridges through libvirt? -- Regards Ananth

2 3