- Users - libvirt List Archives

[libvirt-users] Detecteion of a channel device on a Windows Guest
by Jean-Pierre Ribeauville 07 Sep '15

07 Sep '15

Hi Are there some docs/hint concerning following topic : - After adding a channel device in a Windows 2008 R2 guest .xml conf file : is there anything to do so this device appears "automagically " in the Windows Guest ? Thx. Regards, J.P. Ribeauville P: +33.(0).1.47.17.27.87 Puteaux 3 Etage 5 Bureau 4 jpribeauville(a)axway.com<mailto:jpribeauville@axway.com> http://www.axway.com<http://www.axway.com/> P Pensez à l'environnement avant d'imprimer.

1 0

[libvirt-users] When adding a channel device, we need a shutdown
by Jean-Pierre Ribeauville 07 Sep '15

07 Sep '15

Hi, It looks like , when adding a channel device, that the Guest must be powered off / on to be able to detect the new device . ( simple reboot isn't not enough) Is there any way to avoid this power_cycle ? Thx. J.P. Ribeauville P: +33.(0).1.47.17.27.87 Puteaux 3 Etage 5 Bureau 4 jpribeauville(a)axway.com<mailto:jpribeauville@axway.com> http://www.axway.com<http://www.axway.com/> P Pensez à l'environnement avant d'imprimer.

2 3

[libvirt-users] Can't get PCI passthrough working
by Phill Edwards 06 Sep '15

06 Sep '15

I'm trying to get PCI passthrough working by passing through a DVB-T TV card to a VM running MythTV. When I assign the PCI devices in virt-manager and attempt to start the VM I get this error: Error starting domain: internal error: process exited while connecting to monitor: 2015-09-06T12:41:09.725990Z qemu-kvm: -device vfio-pci,host=05:02.0,id=hostdev0,bus=pci.0,addr=0x9: vfio: error, group 11 is not viable, please ensure all devices within the iommu_group are bound to their vfio bus driver. 2015-09-06T12:41:09.726034Z qemu-kvm: -device vfio-pci,host=05:02.0,id=hostdev0,bus=pci.0,addr=0x9: vfio: failed to get group 11 2015-09-06T12:41:09.726049Z qemu-kvm: -device vfio-pci,host=05:02.0,id=hostdev0,bus=pci.0,addr=0x9: Device initialization failed. 2015-09-06T12:41:09.726064Z qemu-kvm: -device vfio-pci,host=05:02.0,id=hostdev0,bus=pci.0,addr=0x9: Device 'vfio-pci' could not be initialized I've followed a whole load of instructions and made teh settings below: - Updated /etc/sysconfig/grub tp add intel_iommu=on: GRUB_CMDLINE_LINUX="rd.lvm.lv=centos/root rd.lvm.lv=centos/swap rhgb quiet intel_iommu=on" - Allowed unsafe interrupts in /etc/modprobe.d/PME-vfio_iommu_type1.conf: options vfio_iommu_type1 allow_unsafe_interrupts=1 - Blacklisted the PCI card's drivers in /etc/modprobe.d/blacklist.conf so they don't get grabbed by the CentOS hypervisor host: blacklist cx8800 blacklist cx8802 blacklist cx88_dvb blacklist cx88-mpeg Can anyone please advise why I'm getting that error preventing the VM from booting?

1 0

[libvirt-users] Can't get cable connection working on virtual router machine
by Phill Edwards 05 Sep '15

05 Sep '15

I'm pretty new to KVM and have a KVM CentOS 7.1 hypervisor running a few VMs. I'm moving all my VMs from an ESXi host as I want to use KVM in future. Most of my VMs are working except for one which is running a Sophos UTM router (Sophos UTM is similar to products like pfSense <https://www.pfsense.org/>, Smoothwall <http://www.smoothwall.org/> etc). The host has 3 physical NICs which are configured on the Sophos VM as: 1) LAN (fixed IP) 2) DMZ (fixed IP) 3) WAN (which is directly plugged into a cable modem for the internet connection and is configured DHCP). I have imported the settings from the "old" Sophos machine so I know the configuration of the new one is identical to the old one. I have even tried configuring the NICs to have the same MAC addresses as the old one. The problem is that no matter what I try I cannot get the WAN NIC to get an internet link up and running with my cable modem. I have re-installed the VM countless times, turned off the modem and VM, done a factory reset of the modem, and, as I mentioned, ensured the MAC addresses are the same. Nothing I try has been successful. The network interfaces on the new Sophos VM look like this: 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:79:d4:de brd ff:ff:ff:ff:ff:ff inet 192.168.0.254/24 brd 192.168.0.255 scope global eth0 valid_lft forever preferred_lft forever 3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1472 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:79:d4:e8 brd ff:ff:ff:ff:ff:ff 4: eth2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 00:0c:29:79:d4:f2 brd ff:ff:ff:ff:ff:ff inet 192.168.1.254/24 brd 192.168.1.255 scope global eth2 valid_lft forever preferred_lft forever I also found this in the /var/log/system.log of the Sophos VM: 2015:08:29-12:04:05 sop dhclient: DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 6 2015:08:29-12:04:11 sop dhclient: DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 13 2015:08:29-12:04:24 sop dhclient: DHCPDISCOVER on eth1 to 255.255.255.255 port 67 interval 2 2015:08:29-12:04:26 sop dhclient: No DHCPOFFERS received. I have shut down firewalld on the KVM host so I don't think there are any firewall rules blocking this. As soon as I fire up the original Sophos VM on ESXi the internet connection works perfectly again. If I can't get this VM running on KVM it's a show-stopper. Can anyone suggest what might be going on that is preventing the WAN link from connecting? Or suggest a way of troubleshooting this? Thanks in advance.

4 7

[libvirt-users] Error during virt-install, error is not for the guest being instantiated
by DeVincentis, Joseph 04 Sep '15

04 Sep '15

Hi, I have two processes spinning up independent sets of guests, letting the guests run for a while, and tearing them down. The two processes are generally giving their set of guests a different "run-period", so the guest spin-up/tear-down will happen during different cycles/phases. Call these processes A and B, each is currently managing 32 guests. I am seeing an issue, where when process A is spinning guests up, and process B is tearing them down, where on process A, a virt-install will fail, but the error identifies a recently terminated guest which is managed by process B. The guest being instantiated does not get instantiated. "ERROR Error: --network default: Domain not found: no domain with matching uuid 'd39587cb-a2e8-4989-8dce-12a12eb2815e' (process-B-guest-name)" This message is logged by my script at 3:29:23PM. This message is in /var/log/messages; Sep 1 15:29:21 splat04 systemd-machined: Machine qemu-process-B-guest-name terminated. Seems to me that there is some sort of mixup going on, process-B-guest-name is terminated during the virt-install for process-A-guest-name. Any thoughts on this? Like hints on what to look for? This is on Centos7, the version of libvirt is 1.2.8, virt-install is 1.1.0 Thanks - Joe

2 2

[libvirt-users] about guest live migration configuration changes persistence ..
by Jean-Pierre Ribeauville 03 Sep '15

03 Sep '15

Hi, Is a change , by edit it via virsh edit , in the "local" .xml configuration file (i.e. adding a channel device) persistent when live migrating the guest on another host ? Thx. J.P. Ribeauville P: +33.(0).1.47.17.27.87 Puteaux 3 Etage 5 Bureau 4 jpribeauville(a)axway.com<mailto:jpribeauville@axway.com> http://www.axway.com<http://www.axway.com/> P Pensez à l'environnement avant d'imprimer.

3 5

[libvirt-users] New libvirt guide
by David Ashley 02 Sep '15

02 Sep '15

All - Let me introduce myself. I am W. David Ashley, the primary author of the "Libvirt Application Development Guide Using Python" which will soon be published on the libvirt.org web site. I hope all of you will enjoy the new guide when it becomes available. I have multiple decades of experience in writing documentation and training guides and I have about eight years of experience using libvirt and qemu/kvm virtual machines. Most of that experience was in creating and maintaining VM on-demand systems. In the meantime, I am planning a new guide and I would appreciate some feedback from the libvirt user community concerning the potential usefulness and contents of the new guide. The proposed title of the guide will be "Automating Virtual Machines". The current (very) rough outline is: Introduction Intro to virtual machines Installing virtual machines Using virtual machines Using Python to access VMs A sample problem Problem statement Solution requirements Using VMs to solve the problem New problems introduced by using VMs How to programmatically access VMs Intro to the python libvirt module How to connect to a VM How to access and control a guest domain Information the python libvirt module can not provide Storing information about VMs Deciding what information should be stored Using simple text files Using a simple database Securing your information How to set up a VM on demand environment Discovering a VMs ip address Using cron to start processes Starting up a VM Invoking a program on the VM Using SSH to access a VM Shutting down a VM Alternatives to starting/stopping a VM Using the libvirt guest agent Installing the libvirt guest agent Red Hat, Fedora, CentOS openSuse, SuSE Ubuntu, Debian, Mint Using (querying) the libvirt guest agent Logging VM activities Host activities VM activities Securing your VMs Host security VM security Any feedback/suggestions you have will be appreciated and I assure you they will be given serious consideration. At this point, nothing has been written except this rough outline so this is your chance to help form the contents or even make suggestions for a completely different guide. Feel free to post back to this list or send me private email. W. David Ashley w.david.ashley(a)gmail.com

3 2

[libvirt-users] VM locking
by Prof. Dr. Michael Schefczyk 01 Sep '15

01 Sep '15

Dear All, I am trying to use VM (disk) locking on a two node Centos 7 KVM cluster. Unfortunately, I am not successful. Using virtlockd (https://libvirt.org/locking-lockd.html) I get each host to write the zero length file with a hashed filename to the shared folder specified. Regardless of which host I start a VM (domain) on, they do produce the identical filename per VM. What does not work, however, is to prevent the second host to start the VM already running on the first VM. Using sanlock (https://libvirt.org/locking-sanlock.html) no domains do start at all. However, the file "__LIBVIRT__DISKS__" is also not written to the shared folder. Hence, nothing works. Can someone please point me to the right direction? My system is current Centos 7 using a gluster storage setup like suggested for oVirt (https://community.redhat.com/blog/2014/10/up-and-running-with-ovirt-3-5/) based on the oVirt 3.5 repo but without an engine. I do this, because I want to retain VM files in qcow2 format for live backups and with human readable names. What does work then is, e.g., live migration. A locking mechanism to prevent starting a VM twice would be good. Please note that this configuration - both according to redhat and to my own trial and error - requires Lock=False in /etc/nfsmount.conf. Is there a connection with my findings? My issues occur regardless of the files being in NFS or Gluster folders. KVM must load the Gluster storage indirectly via NFS rather than straight via Gluster, as Gluster storage does not seem to fully work, at least in virt-manager. My software versions are: libvirt 1.2.8-16.el7_1.3.x86_64 qemu-kvm ev-2.1.2-23.el7_1.3.1.x86_64 Regards, Michael

2 2

[libvirt-users] Live migration & storage copy broken since 1.2.17
by Antoine Millet 31 Aug '15

31 Aug '15

Hi, It seems that live migration using storage copy is broken since libvirt 1.2.17. Here is the command line used to do the migration using virsh: virsh migrate --live --p2p --persistent --undefinesource --copy -storage-all d2b545d3-db32-48d3-b7fa-f62ff3a7fa18 qemu+tcp://dest/system XML dump of my storage: <pool type='logical'> <name>local</name> <uuid>276bda97-d6c2-4681-bc3f-0c8c221bd1b1</uuid> <capacity unit='bytes'>1024207093760</capacity> <allocation unit='bytes'>4294967296</allocation> <available unit='bytes'>1019912126464</available> <source> <name>hkvm</name> <format type='lvm2'/> </source> <target> <path>/dev/hkvm</path> </target> </pool> When it fails, I get the following error: error: Storage volume not found: no storage vol with matching name '402aea08-26aa-45b6-a46e-a3b02137ff26' Where 402aea08-26aa-45b6-a46e-a3b02137ff26 is the volume bound to my VM: <disk type='volume' device='disk'> <driver name='qemu' type='raw' cache='none'/> <source pool='local' volume='402aea08-26aa-45b6-a46e -a3b02137ff26'/> <backingStore/> <target dev='vda' bus='virtio'/> <alias name='virtio-disk0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </disk> Here is the test result (version is for both source and destination): - 1.2.15 -> ok - 1.2.16 -> ok - 1.2.17 -> fails - 1.2.18 -> fails Is it a bug? Do I miss something? Thanks, Antoine

1 1

[libvirt-users] CPU feature 'svm' not reported inside guest without 'host-passthrough'
by Sladjan Ristic 30 Aug '15

30 Aug '15

Hi, I am confused by the behaviour of the CPU description in my VM. I have a host with an AMD CPU with SVM feature and I want to try nested virtualization on a Fedora22 guest. The host is Fedora22 as well. libvirtd (libvirt) 1.2.13.1 kernel 4.1.6-200.fc22.x86_64 1. I tried 'custom' mode with model qemu64 with required feature 'SVM' but '/proc/cpuinfo' in the guest doesn't show SVM. 2. However, if I forbid the SVM feature, the guest can't be created and libvirt complains that the host CPU has SVM. 3. The only way to have SVM reported in '/proc/cpuinfo' inside the guest VM is using 'host-passthrough' mode, but I want to avoid it to allow migration of the VM. Is this an issue with AMD This is part of the description of the guest: <domain type='kvm'> <name>vagrant-fedora-libvirt-nested_default</name> <uuid>4a379890-b891-4971-aad6-12fad45eaebc</uuid> <memory unit='KiB'>8388608</memory> <currentMemory unit='KiB'>8388608</currentMemory> <vcpu placement='static'>4</vcpu> <os> <type arch='x86_64' machine='pc-i440fx-2.3'>hvm</type> <boot dev='hd'/> </os> <features> <acpi/> <apic/> <pae/> </features> <cpu mode='custom' match='exact'> <model fallback='allow'>qemu64</model> <feature policy='require' name='svm'/> </cpu> Regards, Sladjan

1 0