- Users - libvirt List Archives

[libvirt-users] esx: upload the volume
by Matwey V. Kornilov 14 Aug '14

14 Aug '14

Hi, I am running virsh 1.0.2 with standalone esxi 5.5up1 I would like to upload iso image to the esxi. I am executing commands vol-create followed by vol-upload, that leads me to success for libvirt/qemu. But with esx driver I get the following. virsh # vol-create-as --pool datastore2 --name coreos_production_iso_image.iso --format raw --capacity 139460608 error: Failed to create vol coreos_production_iso_image.iso error: internal error Volume name 'coreos_production_iso_image.iso' doesn't have expected format '<directory>/<file>' virsh # vol-create-as --pool datastore2 --name 'coreos/coreos_production_iso_image.iso' --format raw --capacity 139460608 error: Failed to create vol coreos/coreos_production_iso_image.iso error: internal error Volume name 'coreos/coreos_production_iso_image.iso' has unsupported suffix, expecting '.vmdk' The error messages lead me to idea that I am doing something wrong.

1 0

[libvirt-users] Behavior of disk caching with qcow2 disks
by Andrew Martin 14 Aug '14

14 Aug '14

Hello, I am running several virtualization servers with QEMU 1.4.x and libvirt 1.0.2 on Ubuntu 12.04 and am working on optimizing the cache= and aio= options for the virtual machines. These VM images are mostly qcow2, and are served both from a local ext4 filesystem (with data=ordered,barrier) and from an NFS mountpoint (with sync). The local filesystem sits on top of an md software RAID of SATA HDDs. I have read some conflicting information about which cache option is used by default. This documentation states that cache=writethrough is the default: http://www-01.ibm.com/support/knowledgecenter/linuxonibm/liaat/liaatbpkvmgu… However this SuSE documentation claims that QEMU 1.2.x and newer allows the driver to select which cache mode, and it often defaults to cache=writeback: https://www.suse.com/documentation/sles11/book_kvm/data/sect1_1_chapter_boo… Which is correct? How is the cache mode set by default (if cache= is not specified)? My second question is can cache=none be used safely on a local ext4 filesystem with no BBU? Since ext4 uses barriers, would writing to these qcow2 image files be safe? The kernel documentation about barriers states that "Write barriers enforce proper on-disk ordering of journal commits, making volatile disk write caches safe to use, at some performance penalty". Does this apply to qcow2 VM images? Thanks, Andrew Martin

2 2

[libvirt-users] Error when compiling libvirt 1.2.7 on CentOS 6.4
by Hongbin Lu 14 Aug '14

14 Aug '14

Hi, I was trying to follow this guide (http://libvirt.org/compiling.html) to compile libvirt1.2.7 on CentOS6.4. There is an error on executing "make". Below is the steps to reproduce and the error messages. Any help is appreciated. Thanks. $ sudo yum install gcc make gnutls-devel device-mapper-devel \ python-devel libnl-devel yajl-devel \ libxml2-devel libpciaccess-devel $ wget ftp://libvirt.org/libvirt/libvirt-1.2.7.tar.gz $ gunzip -c libvirt-1.2.7.tar.gz | tar xvf - $ cd libvirt-1.2.7/ $ ./configure --prefix=/usr --localstatedir=/var --sysconfdir=/etc --with-openvz=yes $ make ..... Making all in src make[2]: Entering directory `/home/ec2-user/libvirt-1.2.7/src' GEN libvirt_access.xml /bin/sh: -w: command not found make[2]: *** [libvirt_access.xml] Error 127 make[2]: Leaving directory `/home/ec2-user/libvirt-1.2.7/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/ec2-user/libvirt-1.2.7' make: *** [all] Error 2 $ make V=1 .......... /usr/bin/ld:libvirt.syms:3220: syntax error in VERSION script collect2: ld returned 1 exit status make[3]: *** [libvirt.la] Error 1 make[3]: Leaving directory `/home/ec2-user/libvirt-1.2.7/src' make[2]: *** [all] Error 2 make[2]: Leaving directory `/home/ec2-user/libvirt-1.2.7/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/ec2-user/libvirt-1.2.7' make: *** [all] Error 2

3 4

[libvirt-users] provide http server path as vfat floppy/disk
by Vasiliy Tolstov 13 Aug '14

13 Aug '14

Hello. Is that possible to use some url (http/https) that contains files as source for floppy or block device? I can find only http://www.redhat.com/archives/libvir-list/2009-November/msg00604.html that says about local files and floppy, i can write/use fuse based fs , but may be libvirt/qemu have this internally? -- Vasiliy Tolstov, e-mail: v.tolstov(a)selfip.ru jabber: vase(a)selfip.ru

1 0

[libvirt-users] Support live migration of OpenVZ container
by Hongbin Lu 13 Aug '14

13 Aug '14

Hi libvirt team, I was trying to migrate an OpenVZ domain from one host to another host. It seems that the live migration of OpenVZ container is there ( https://openvz.org/Migration_from_one_HN_to_another) but it haven't been integrated to libvirt yet. In particular, I ran the following commands and it failed: $ sudo virsh migrate --live --verbose --undefinesource 101 openvz+ssh:// root(a)10.12.1.39/system error: this function is not supported by the connection driver: virDomainMigrate3 Could anyone confirm if libvirt supports live migration of OpenVZ domain. If yes, please let me know the correct way to do that. If no, I would like to know if there is a plan to add support for that. I don't mind to submit a patch if your team likes to review it. Please let me know. Thanks. Best regards, Hongbin

2 1

[libvirt-users] how virt-manager gets CPU usage of Guest
by Restituto Marcus Arevalo 13 Aug '14

13 Aug '14

Hi, Does anybody know how virt-manager is getting the CPU usage of the guests? I need to know how to get the CPU usage of the guests and this is the only way I can think of. In other words, how does virt-manager produce the graph of the CPU usage of EACH guest in KVM? I don't actually need the graph, I just need to get the percentage(%) since I'm trying to create a Java program to get the CPU usage of each guest in KVM. Thanks! regards, Marco

2 2

[libvirt-users] Does libvirt have API to detect image format
by Jianfeng Tang 13 Aug '14

13 Aug '14

> I am using backing store when I create my volume. However, the backing store file is from another team, which can be either in raw or in qcow2 format. Is there a libvirt API to detect image format by given an image file? Or I have to run qemu-img? >

3 4

[libvirt-users] [qemu-kvm] Network bandwidth limits via libvirt
by Romain Hardouin 13 Aug '14

13 Aug '14

Hi all, I try to understand why the interface bandwidth limits are not applied to my Ubuntu 12.04 VM (qemu-kvm driver) and I would need some insights. I spawn my VM with OpenStack and when I issue a dumpxml on my domain, I can see that bandwidth limits are set: <interface type='ethernet'> <mac address='02:68:b5:c6:fa:a5'/> <bandwidth> <inbound average='500' peak='500' burst='1000'/> <outbound average='500' peak='500'/> </bandwidth> <script path=''/> <target dev='tap68b5c6fa-a5'/> <model type='virtio'/> <alias name='net0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> However when I send a big file with scp I don't see any limits applied to the transfer rate. I've checked the domain log and I didn't see any qemu command line argument related to bandwidth limits. Should I expect to see a specific argument like "bps_wr" for IO limits? Also, should I expect to see something in particular under the cgroup directory for the qemu instance? Or do network limits are managed in another way than with cgroups? Below some information about my hypervisor: % virsh -V Virsh command line tool of libvirt 1.2.2 See web site at http://libvirt.org/ Compiled with support for: Hypervisors: QEMU/KVM LXC UML Xen LibXL OpenVZ VMWare VirtualBox Test Networking: Remote Network Bridging Interface netcf Nwfilter VirtualPort Storage: Dir Disk Filesystem SCSI Multipath iSCSI LVM RBD Sheepdog Miscellaneous: Daemon Nodedev AppArmor Secrets Debug Readline Modular qemu-kvm: 2.0.0 Kernel: 3.2.0-64-generic Any hints will be appreciated :) Thanks, Romain

2 1

[libvirt-users] IP/MAC antispoof-protection
by Anton Gorlov 11 Aug '14

11 Aug '14

Hi all. What right way to protect ip/mac spoofing for guests withnount dhcp and other 1 ip per guest?

2 1

[libvirt-users] Passing literal -cpu model string to qemu
by Richard W.M. Jones 11 Aug '14

11 Aug '14

On aarch64 with -M virt, the default CPU model is cortex-a15 (a 32 bit CPU). This is IMHO a stupid default, but there we are. Therefore most users will need to pass the `-cpu cortex-a53' or `-cpu cortex-a57' flag to qemu, depending on a complex formula of their host CPU and if they are using TCG or not. However I cannot work out how to pass this through libvirt. The obvious one would be: <cpu><model>cortex-a57</model></cpu> This passes `-cpu host' which is both wrong and likely to be a bug. I tried adding various flags and studying the source, but it's extremely non-obvious how to do this. Also adding the following into cpu_map.xml didn't make any difference: --- a/src/cpu/cpu_map.xml +++ b/src/cpu/cpu_map.xml @@ -626,6 +626,11 @@ <vendor name='IBM'/> <pvr value='0x004b0100'/> </model> + </arch> +  + <arch name='aarch64'> + <model name='cortex-a53'/> + <model name='cortex-a57'/> </arch> </cpus> Ideas here? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://people.redhat.com/~rjones/virt-df/

2 1

[libvirt-users] Windows 8 virtualized on a Linux host
by arnaud gaboury 09 Aug '14

09 Aug '14

I plan to dual boot w8 and Archlinux, and virtualize the same w8 machine with libvirt qemu/kvm. I already did the dual boot. I am now starting the virtualization. My plan was to use the w8 partition as a Disk volume pool [1]. Unfortunately, I didn't noticed only FAT32 was a valid format when I installed w8 on NTFS. I need to change my whole setup, but beforethat, I want to be sure libvirt does not accept NTFS. So please, can you confirm? Then, it seems I am in trouble as my readings let me think w8 will not install on FAT32. Am I correct? Now it sounds to me I am in a dead end. Does anyone have any hint to achieve my plan? Thank you for any suggestions. [1]http://libvirt.org/storage.html

2 1

[libvirt-users] Live blockcopy onto storage pool that is an NFS mount?
by Andrew Martin 08 Aug '14

08 Aug '14

Hello, I am running qemu-kvm 1.4.0 and libvirt 1.0.2 on Ubuntu 12.04. I have two NFS mountpoints configured as two separate pools in virsh: <pool type='dir'> <name>nfs1</name> <uuid>419d799c-2493-6ebc-6848-53b0919e7bad</uuid> <capacity unit='bytes'>6836057014272</capacity> <allocation unit='bytes'>0</allocation> <available unit='bytes'>6836057014272</available> <source> </source> <target> <path>/var/lib/libvirt/images/nfs1</path> <permissions> <mode>0711</mode> <owner>0</owner> <group>0</group> </permissions> </target> </pool> The XML for the nfs2 pool is similar, with the mountpoint being /var/lib/libvirt/images/nfs2. I would like to be able to move the VM disk from one of the NFS mountpoints to another (nfs1 --> nfs2) live. I followed the instructions here: http://thr3ads.net/libvirt-users/2013/02/2174486-Migrate-VMs-to-different-s… However when I attempt the blockcopy command, I receive this error: Failed to create file '/var/lib/libvirt/images/nfs2/vm.qcow2': Operation not permitted Is it not possible to blockcopy in this way to an NFS mount? Thanks, Andrew

2 2

[libvirt-users] File permissions on VNC unix sockets
by Landon Gilbert-Bland 07 Aug '14

07 Aug '14

I am using kvm libvirt/qemu, and am trying to get VNC working with unix sockets. I'm using the following in my XML: <graphics type='vnc' socket='/tmp/kvmtest'/> This works, it creates the unix socket and I can use it for VNC. But it creates the socket with 755 permissions, and owned by libvirt-qemu:kvm. Ideally, I would like it to be 770 root:libvirtd, but could also work with 775 libvirt-qemu:kvm. Basically, I would like the group to be read/write/execute, so that anyone in this group can use virt-manager (or whatever) to get into the vm with vnc. I haven't found a way to change this in the settins. It doesn't seem to be honoring the unix socket settings in /etc/libvirt/libvirdd.conf (it looks like those are only for libvirt-sock and libvirt-sock-ro), and I haven't been able to find any documentation about changing these permissions in the XML. I can manually chmod/chown the socket after it is created (manually or with a cron), but that is far from an ideal solution. I don't suppose anyone has run across a configuration option for this that I have missed? I have only tested this in ubuntu 14.04.1 and debian testing, for what it's worth. Thanks, --landon

1 0

[libvirt-users] Understanding Libvirt Java API in order to get CPU usage of a VM
by Restituto Marcus Arevalo 07 Aug '14

07 Aug '14

I'm having a problemin understanding the Libvirt Java API. Here is my sourcein reading some of the classes: http://libvirt.org/sources/java/javadoc/ I have no idea where to start in order for me to get the CPU usage. Can anybody help? Or at least give me some points on where to start in order for me to get the CPU usage of a VM in KVM? Any suggestions will really help. Thanks regards, Marco

2 1

[libvirt-users] [help] Does "virsh blockpull" works on live virtual machine
by chenyanqiu＠keytonecloud.com 07 Aug '14

07 Aug '14

Hi all, I have a kvm virtual machine running (qemu version 2.0), and I had took several external snapshots of the disk( using "virsh snapshot-create-as"). Now, the existed disk files relationship look like: base <- snap1 <-snap2 <- current using disk file. Now I want to remove snap1 and snap2, and let current disk using the base image file as back file directly. Unfortunately, for some reason, I can not shutdown the vm, and exectue the "virsh blockpull" command. Can I execute the "virsh blockpull" while the vm is power on, and the application in the vm is not affected? Thanks very much! Daniel

2 2

[libvirt-users] possible to dynamically set/add network domain name?
by Rich Megginson 06 Aug '14

06 Aug '14

trying to use virsh net-update to add a domain name libvirt-1.1.3.5-2.fc20.x86_64 sudo virsh net-update default add domain "<domain name='test'/>" --live --config error: this function is not supported by the connection driver: can't update 'domain' section of network 'default' is it possible to add a domain using net-update?

2 2

[libvirt-users] username and passwd after create VM from XML dump
by Yuanzhen Gu 06 Aug '14

06 Aug '14

Good afternoon guys, I created the VM from XML dump, # virsh create vm4.xml and # virsh list I can see the VM running, and I know the ip address for the VM, but how can I know the username and password to login the running VM, which is created from XML dump? is there any place to edit and add the username and password in guest domain XML ? thanks a lot!! Best, Yuanzhen

2 6

[libvirt-users] Ubuntu Trusty: failed to create VM due to permission denied
by Jianfeng Tang 06 Aug '14

06 Aug '14

> Hi, libvirt experts, I used libvirt to create a VM and used backing store to a local file. It works fine until I installed my box to Unbuntu Trusty (14.04). I got the following errors when I tried to start the VM: Could not open backing file: Could not open <path to my backing file>: Permission denied However, if I moved my image file (not base image) to default location /var/lib/libvirt/images. It works. It seems something related to selinux. However, my box seems not have selinux installed at all. Anyone knows how to fix this? Thanks, ~Frank >

2 4

[libvirt-users] libvirt external disk-only snapshot will pause the VM?
by Yuanzhen Gu 06 Aug '14

06 Aug '14

Hello folks, If I give command to take external disk-only live snapshot, such as # virsh snapshot-create-as vm1 snapshot_vm1 --disk-only will the guest OS be paused or suspend while the snapshot being taken? Thanks very much! Best, Yuanzhen

2 8

[libvirt-users] error configuring a public bridge --> uvt-kvm: error: libvirt domain 'newtest' has no NIC MACs available.
by Jordi Arjona 06 Aug '14

06 Aug '14

Hi all, First at all, excuse if this is not the right place for sending this email. I'm using ubuntu 14.04 server edition and i'm trying to deploy some virtual machines in a server in such a way that they are visible in the same network that the host machine. The tools I'm using are the uvtool package and libvirt. The error I'm getting is the following: "uvt-kvm: error: libvirt domain 'newtest' has no NIC MACs available." Here I'm providing the quickest way (I think) of reproducing my problem. The steps to follow would be: 1) Creating a bridge. This is my /etc/network/interfaces # This file describes the network interfaces available on your system # and how to activate them. For more information, see interfaces(5). # The loopback network interface auto lo iface lo inet loopback auto em1 iface em1 inet manual auto br0 iface br0 inet static address 9.109.124.62 network 9.109.124.0 netmask 255.255.255.0 broadcast 9.109.124.255 gateway 9.109.124.1 #iface br0 inet dhcp bridge_ports em1 bridge_stp off bridge_fd 0 bridge_maxwait 0 dns-search xx.xx.xx.xx dns-nameservers 192.168.122.1 9.0.146.50 9.0.148.50 2) downloading a trusty tahr cloud image and run it uvt-simplestreams-libvirt sync release=precise arch=amd64 uvt-kvm create test release=trusty 3) duplicate the xml generated when running the machine in order to modify the network interface by the bridge one. In the XML I also remove the uuid and change the VM name. sudo mv /etc/libvirt/qemu/test.xml /etc/libvirt/qemu/test2.xml In test 2 the most relevant change is replacing <interface type='network'> <mac address='52:54:00:86:7a:31'/> <source network='default'/> <model type='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> by <interface type='bridge'> <source bridge='br0'/> <model type='virtio'/> </interface> 4) Once I have my new XML, I define and start the new machine uvt-kvm destroy test virsh undefine test sudo virsh define /etc/libvirt/qemu/test2.xml virsh start newtest (newtest is the name provided in test2.xml) 5) Up to this point everything seems to be fine, but if i run a wait command i get the error uvt-kvm wait newtest --insecure uvt-kvm: error: libvirt domain 'newtest' has no NIC MACs available. I have checked the xml generated when defining newtest and there is a MAC in there, so it is not a problem generating a MAC address. Also the interface br0 is visible for virsh: virsh iface-list Name State MAC Address --------------------------------------------------- br0 active xx:xx:xx:xx:xx:xx lo active 00:00:00:00:00:00 I have been googling for a while and couldn't find anything but this: http://pastebin.com/em0hmy4t That tries something similar and falls in the same error. Any help will be very much appreciated. It is the first time i send anything to one of this mail lists, so, please, if you need me to provide any other info, tell me. Thanks! Jordi

1 0

[libvirt-users] How to import libvirt library in Java
by Restituto Marcus Arevalo 05 Aug '14

05 Aug '14

Hi, I am trying to get the CPU/Memory usage of the VMs in KVM using Java. But before I get to start coding, I have no idea where to get the libvirt library and how to import it in Java. Does anyone know where I can get it and how I'll be able to use its library? Thanks regards, Marco

1 0

Re: [libvirt-users] [libvirt] help
by Michal Privoznik 05 Aug '14

05 Aug '14

[Dropping libvirt-announce list] On 05.08.2014 11:47, Himanshu Sharma wrote: > Dear Team, > Greetings!! > > I'm testing libvirt with VMware ESXi. I'm able to connect to VMware ESXi with libvirt driver but not able to connect it through "Virtual Machine manager" GUI also not able to run virt-clone command inturn to clone ESXi VM. Can you please guide me how to do so? > The virt-manager does not support ESX yet. It merely supports QEMU/KVM, XEN and LXC. And for the virt-clone issue: 1) "computer broken" is usually not much descriptive. Is there any error message you are seeing? What behavior are you experiencing? 2) virt tools have a separate mailing list: http://www.redhat.com/mailman/listinfo/virt-tools-list Try asking there. Usually it's considered bad practice to repeat your question every 5 minutes, crosspost it over several mailing lists (even libvirt-announce which is not user support channel but a low traffic list for announcements as it name suggests) or forward the e-mail that is irrelevant (Welcome to the list). Michal

1 0

[libvirt-users] libvirt and lvm thin pool
by Vasiliy Tolstov 04 Aug '14

04 Aug '14

Hi all. I'm using libvirt 1.2.6 I want to use lvm storage for my virtual machines. But i want to use new lvm2 feature - thin pool. How can i do that in libvirt? If libvirt can't create it via pool xml, does (and how) to use this setup under libvirt? -- Vasiliy Tolstov, e-mail: v.tolstov(a)selfip.ru jabber: vase(a)selfip.ru

3 3

[libvirt-users] "virsh list" doesn't show correct state in case of no reply from Guest OS
by Koji Inami 04 Aug '14

04 Aug '14

Hello, we were facing a bit of problem that "virsh list" command didn't show correct state, and I report it. I'm talking about a following message, for example. [root@server ~]# virsh list Id Name State ---------------------------------------------------- 1 rhel65 running The command showed "running" in spite of no reply from a Guest OS, while qemu process was dumping a corefile. So I guess the process was dying at that time. I'm sure the command shouldn't show "running" state because the process was not running. I believe that "crushed" or "shut off" are better than "running" in virDomainState. Please see: http://libvirt.org/html/libvirt-libvirt.html I'd like to hear your thinking. Best Regards, Koji

1 0

[libvirt-users] virtio hotplug in CentOS 5.10 guest incomplete
by Dennis Jacobfeuerborn 01 Aug '14

01 Aug '14

Hi, I just tried to hot-plug a virtio disk to a CentOS 5.10 guest that runs on a CentOS 6.5 host. The Problem is that while the device shows up under /sys/ the corresponding device node does not: /sys/devices/virtio-pci/virtio0: bus device driver modalias net:eth0 power status subsystem uevent vendor /sys/devices/virtio-pci/virtio1: block:vda bus device driver modalias power serial status subsystem uevent vendor /sys/devices/virtio-pci/virtio2: bus device driver modalias power status subsystem uevent vendor So virtio0 is eth0 and virt1 is vda but virtio2 has no device associated with it. Any ideas why the system doesn't pick up the device all the way? Regards, Dennis

1 0