- Users - libvirt List Archives

[libvirt-users] With no luck with virDomainGetInfo and virDomainMemoryStats for memory usage of a running vm
by 张章 28 Feb '13

28 Feb '13

Hi, all I am trying to get used memory of a running vm using libvirt and then to calculate the memory usage, but with no luck. 1.The used memory returned by virDomainGetInfo is equal to max memory with no virtio balloon driver set. When using virtio balloon driver and setting currentMemory less than memory in the xml, the used memory returned by virDomainGetInfo is nearly equal to currentMemroy. Again, I can't get the memory used by vm. 2. When I turned to virDomainMemoryStats, I can only get two values : Current balloon value and Resident Set Size of the process running the domain. (for kvm 0.14.1 and libvirt 1.0.2) That's not what I want. I only need to know how much memory used by the running vm, so I can calculate the memory usage of this vm. (like the used memory return by cmd "free -m") Any hints? Many thanks! Z. ZhangNational Research Center for intelligent Computing Systems,ICT, Chinese Academy of Sciences

1 0

[libvirt-users] Deleting and coalescing live snapshots
by Skardal, Harald 28 Feb '13

28 Feb '13

All, I have a service that takes new live KVM snapshots Si regularly, keeps a fixed number N (Si ,..,Si-N+1), and therefore needs to delete Si-N in this cycle. Until libvirt includes support for this capability that is said to be available in qemu, what is a safe workflow to delete old live KVM snapshots w/o losing data. Do I need to pause/shut down the VM? The development environment is Fedora 18 with qemu, libvirtd and libvirt upgrades to the more recent stable versions. Harald Skardal, Stratus Technologies.

3 2

[libvirt-users] The accurate CPU usage of a domain?
by 张章 28 Feb '13

28 Feb '13

Hi,all I want to get a relatively accurate cpu usage of a domain. I have a few questions about virDomainGetInfo: struct virDomainInfo{ unsigned char state : //the running state, one of virDomainState unsigned long maxMem :// the maximum memory in KBytes allowed unsigned long memory :// the memory in KBytes used by the domain unsigned short nrVirtCpu :// the number of virtual CPUs for the domain unsigned long long cpuTime : //the CPU time used in nanoseconds } If a domain is assigned to 4 vcpus and my machine has 6 physical cores, then the cpuTime in virDomainInfo should be the sum of cpu time of the 4 vcpus. When I want to calculate the cpu usage, which of the following is more accurate? 1/4 * cpu time / total time or 4/6 * cpu time / total time or just cpu time / total time regardszhangzhang

2 1

[libvirt-users] error when call migrate function "error: invalid argument: qemuDomainMigratePrepare3: unsupported flags (0x200)"
by loic.cloatre＠orange.com 27 Feb '13

27 Feb '13

Hello, I try to migrate a VM and I get this error : loic@loic-VirtualBox:~/workspace/src/build/bin$ ./virsh --connect qemu:///system migrate --live vmUbuntuPartage2CacheWriteBack qemu+ssh://192.168.10x.2x/system --unsafe loic(a)192.168.10x.2x's password: error: invalid argument: qemuDomainMigratePrepare3: unsupported flags (0x200) Does somebody know this error? I am working on windows 7, I have: 1) Virtual box with two vms ubuntu desktop 12.04: "VM1" and "VM40" 2) One shared hard drive build with virtual box: "newDiskSharedFev.vdi" In my VM1: 1) I have build one "vmUbuntuPartage" ubuntu server 12.10 on shared disk with this command: sudo virt-install -n vmUbuntuPartage -r 1500 --disk path=/media/partage/newDiskSharedFev.vdi,bus=virtio,size=4,cache=writeback -c /home/loic/Downloads/ubuntu-12.04.1-server-i386.iso --network network=default,model=virtio --connect=qemu:///system --vnc -v 2) I have download and built libvirt-0.10.0 (I can give configure command if necessary) In my VM40: 1) I install libvirt with apt-get 2) I have mounted shared folder with "newDiskSharedFev.vdi" So, in my VM1 I try to migrate my "vmUbuntuPartage" from VM1 to VM40 and I get error above. Here my configuration file /etc/libvirt/qemu/ vmUbuntuPartage.xml --> I try to change <driver name='qemu' type='raw' cache='none or writeback or writetrough'/> ================  <domain type='qemu'> <name>vmUbuntuPartage</name> <uuid>5d38cf86-d9c5-b0f2-cc03-982a1f18bede</uuid> <memory>1536000</memory> <currentMemory>1536000</currentMemory> <vcpu>1</vcpu> <os> <type arch='i686' machine='pc-1.0'>hvm</type> <boot dev='hd'/> </os> <features> <acpi/> <apic/> <pae/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/bin/qemu-system-x86_64</emulator> <disk type='file' device='disk'> <driver name='qemu' type='raw' cache='none'/> <source file='/media/partage/SharedDiskUbuntu.vdi'/> <target dev='vda' bus='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </disk> <disk type='block' device='cdrom'> <driver name='qemu' type='raw'/> <target dev='hdc' bus='ide'/> <readonly/> <address type='drive' controller='0' bus='1' unit='0'/> </disk> <controller type='ide' index='0'> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/> </controller> <interface type='network'> <mac address='52:54:00:8b:b1:80'/> <source network='default'/> <model type='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> <serial type='pty'> <target port='0'/> </serial> <console type='pty'> <target type='serial' port='0'/> </console> <input type='mouse' bus='ps2'/> <graphics type='vnc' port='-1' autoport='yes'/> <video> <model type='cirrus' vram='9216' heads='1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <memballoon model='virtio'> <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> </memballoon> </devices> </domain> ================== Loïc Cloâtre _________________________________________________________________________________________________________________________ Ce message et ses pieces jointes peuvent contenir des informations confidentielles ou privilegiees et ne doivent donc pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce message par erreur, veuillez le signaler a l'expediteur et le detruire ainsi que les pieces jointes. Les messages electroniques etant susceptibles d'alteration, France Telecom - Orange decline toute responsabilite si ce message a ete altere, deforme ou falsifie. Merci. This message and its attachments may contain confidential or privileged information that may be protected by law; they should not be distributed, used or copied without authorisation. If you have received this email in error, please notify the sender and delete this message and its attachments. As emails may be altered, France Telecom - Orange is not liable for messages that have been modified, changed or falsified. Thank you.

3 14

[libvirt-users] questions of libvirt's lxc
by Han Yuejuan-B42073 27 Feb '13

27 Feb '13

Hi all, I am trying to init lxc by libvirt instead of using lxc tools. I have several questions as following: software platform: integrating libvirt-0.10.1 to yocto based on sdk1.3 hardwatre platform: p4080ds steps: root@p4080ds:~# cat > container.xml <<EOF <domain type='lxc'> <name>container</name> <memory>500000</memory> <os> <type>exe</type> <init>/bin/sh</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/libexec/libvirt_lxc</emulator> <interface type='network'> <source network='default'/> </interface> <console type='pty' /> </devices> </domain> <EOF root@p4080ds:~# libvirtd -d root@p4080ds:~# virsh --connect lxc:/// define container.xml Domain container defined from container.xml root@p4080ds:~# virsh --connect lxc:/// start container error: Failed to start domain container error: internal error Network 'default' is not active. To resolve this issue I change the container.xml to as below: <domain type='lxc'> <name>container</name> <memory>500000</memory> <os> <type>exe</type> <init>/bin/sh</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/libexec/libvirt_lxc</emulator> <interface type='bridge'> <source network='br0'/> <mac address='00:11:22:34:34:34'/> (I am not sure it will make difference) </interface> <console type='pty' /> </devices> </domain> then error infos as below: root@p4080ds:~# virsh --connect lxc:/// define container.xml error: Failed to define domain from container.xml error: internal error No <source> 'bridge' attribute specified with <interface type='bridge'/> Q1: What an appropriate virtual network is supported by virsh for error mentioned above? Linux container support several types of network such as veth, vlan, macvlan, phys and how to edit different xml file for different network type? Q2: There are two kinds of containers , application container and system container. If using lxc tools, lxc-execute will establish an application container and lxc-create with lxc-start will establish a system container. So with libvirt how to set up these two kinds of containers? by different xml file or by different virsh cmds? Q3: If I want to define a rootfs (such as busybox) for container how I edit the relative xml file? anyone has related infos that can be shared? Q4:Cgroups manage and control resources using for lxc containers. If use libvirt how to realise the similar resource tuning for containers based on PowerPC? There are limited infos at: https://www.berrange.com/posts/2009/12/03/using-cgroups-with-libvirt-and-lx… Regards, Yuejuan

1 0

[libvirt-users] VMs XML difinitions can't be recognized after upgrading to libvirt 0.9.13
by Marwan Tanager 27 Feb '13

27 Feb '13

I used to run many VMs on my Ubuntu 10.04 box. However, after upgrading to Ubuntu 12.10 (and thus, libvirt 0.9.13), The VMs can't be recognized anymore. Nothing appears in the output of 'virsh list --all', nor in virt-manager. When I created a new VM based on a disk image of one of the not-recognized-any-more VMs, a new schema file has been auto-generated, and the new VM now is recognized. So, if this is caused by incompatibility with the old schemas, would the solution be to manually recreate all the VMs by importing their disk images to the newly created VMs, or is there a more intelligent approach? Thanks for your time. Marwan

2 2

[libvirt-users] [libvirt-php] network string name
by Yusof Khalid - FreeBSD / OpenBSD 27 Feb '13

27 Feb '13

Hi list, So this is my first question on this list :) I'm working for my script to get network information for our guest os (Using Xen 3) using PHP libvirt. Just get confused for function libvirt_network_get($res, $name) $name should be network name string as per the doc. What is the value for that ? is it the name in ifconfig list that attach with the guest os e.g vif101 ? Thanks in advance. -- -- 7.2-RELEASE-p6

1 0

[libvirt-users] Routed network w/o libvirt adding iptables rules
by Nils Toedtmann 27 Feb '13

27 Feb '13

Hello In the case of a "routed" libvirt network, is there a way to prevent libvirt from adding *any* iptables rules? Or at least to stop libvirt from inserting its rules at the *top* of the chains (which renders previously inserted rules useless)? I have VPN tunnels connecting VMs with each other and with clients, and the default rules generated by libvirt are not right for my use case. The firewall rulesets i actually need are quite simple. I am very happy to manage them outside libvirt. I am thinking of replacing /sbin/iptables with a dummy, but that is a last ressort which i hope is not necessary. Please advise /nils. PS: See also long standing libvirt issues: * https://bugzilla.redhat.com/show_bug.cgi?id=533193 * https://bugzilla.redhat.com/show_bug.cgi?id=689377

2 1

[libvirt-users] How to add static routes?
by Nils Toedtmann 26 Feb '13

26 Feb '13

Hello i have several libvirt hosts (CentOS6, kvm), each with an internal libvirt "default" network with "forward mode='route'". I have two questions concerning static routes. To give some context: On each of my libvirt hosts there is a VM acting as a OpenVPN-gateway to the other libvirt hosts. [A] How can i add a persistant static route on the libvirt host to one of its guests? Someone had already asked this on this list in 2011 but apperently did not receive an answer [1]. Likewise, i found that a /etc/sysconfig/network-scripts/route-virbr0 does not get applied when libvirt starts its default network. [B] How can i make libvirt's dnsmasq push static routes into the guests? Apperently dnsmasq supports pushing of static routes using DHCP options 33 or 121 [2], e.g. "--dhcp-option=121,x.x.x.x/yy,z.z.z.z". But i did not find how to add that to a libvirt network declaration. I know that if i had [A] working, the host could redirect the IP traffic on behalf of the guests, but i'd like to avoid the resulting extra hop and the flood of ICMP redirects. Please advise. Cheers /nils. [1] https://www.redhat.com/archives/libvirt-users/2011-December/msg00061.html [2] http://tools.ietf.org/html/rfc3442

1 1

[libvirt-users] shutdown a windows guest take ages..... (20- 34 minutes here)
by weber＠zackbummfertig.de 26 Feb '13

26 Feb '13

hi, i use libvirt 1.0.2 (-r1 , gentoo linux). when i create a vm with win7 guest, virtio nic, virtio hdd, all is running fine. but when i shutdown the windows guest, it takes somestimes 20-34 minutes !!! iotop shows me in the whole time writing with 2,xxmb/sec. is the complete machine "rewritten" to disk? i use disk.images on a raid6, but hey, dd shows me it can writes there with 26mb/sec. any ideas what causes these long shutdown times? thanks for any hints, marko

3 3