January 2017 - Users - libvirt List Archives

[libvirt-users] gluster store and autostart - but fails
by lejeczek 23 Jan '17

23 Jan '17

hi all I've a few guest which work/run perfectly fine, I believe, except for autostart. Configuration of system, gluster and libvirt is pretty regular and not complex. Errors I see: ... failed to initialize gluster connection (src=0x7f9424266350 priv=0x7f94242922b0): Transport endpoint is internal error: Failed to autostart VM 'rhel-work2': failed to initialize gluster connection (src=0x7f9 failed to initialize gluster connection (src=0x7f942423fef0 priv=0x7f9424256320): Transport endpoint is internal error: Failed to autostart VM 'rhel-work3': failed to initialize gluster connection (src=0x7f9 failed to initialize gluster connection (src=0x7f9424261b20 priv=0x7f94242a18b0): Transport endpoint is internal error: Failed to autostart VM 'rhel-work1': failed to initialize gluster connection (src=0x7f9 ... I tried to make systemd libvirtd to wait for gluster: After=glusterd.service but if that's all required then, well, still fails. Would you have any suggestions? Many thanks, L.

2 1

[libvirt-users] Controlling the name of the 'tap0' device, in a bridged networking setup
by Govert 21 Jan '17

21 Jan '17

Hi, I'm trying to control the name of the 'tap0' device that gets created as I start a domain that uses bridged networking. The XML specification of the domain contains the following configuration <interface type='bridge'> <source bridge='br0'/> </interface> The libvirt documentation ( http://libvirt.org/formatdomain.html#elementsNICSBridge) and other discussions online tell me that I just need to include the <target dev='desired_dev_name'/> tag in the XML specification of the domain under the <interface> tag. Unfortunately doing so appears to have no effect; the tun device created and 'enslaved' in the bridge is still called 'tap0'. Interestingly, I never get a tun device with a name prefixed by 'vnet' or 'vif' which, according to the documentation, is the default behaviour (?). The host is running CentOS 7, and virsh is used to start the domain. Best, Govert

2 4

[libvirt-users] io=native & io=threads
by W Kern 17 Jan '17

17 Jan '17

Googling provides lots of interesting info on the use of these in various situations such as SSD, number of VMs in the pool etc. What is the default in Libvirt (or is the default 'neither') Sincerely W Kern

2 2

[libvirt-users] Regarding Migration Statistics
by Anubhav Guleria 14 Jan '17

14 Jan '17

Greetings, I am writing a code using libvirt API to migrate VM between two physical hosts *(QEMU/KVM) , *say some *n *number of times. *1)* I am using right now* virDomainPtr virDomainMigrate (.......) *and to calculate the total migration time I am using something like this: *clock_gettime(CLOCK_MONOTONIC_RAW,&begin); * *migrate*(domainToMigrate,nodeToMigrate); *clock_gettime(CLOCK_MONOTONIC_RAW,&end);* *Total Migration Time = end.tv_sec-begin.tv_sec* Is this correct way to calculate total migration time. And is there some way to calculate the downtime (not how to set it)? *2) *I am interested in identifying in particular other statistics of migration like : *Number of iterations in Pre Copy*, *Memory transferred in each iteration* etc. I was going through the API and found* virDomainJobInfo <http://libvirt.org/html/libvirt-libvirt-domain.html#virDomainJobInfo> and virDomainGetJobStats <http://libvirt.org/html/libvirt-libvirt-domain.html#virDomainGetJobStats> *functions.But how to use them is not very clear. Can anyone point me to right place to achieve this objective? Thanks in advance. And sorry if that was too silly to ask. Anubhav

3 3

[libvirt-users] changing to cache='none' on the fly?
by W Kern 10 Jan '17

10 Jan '17

As is already well documented, we find that we need cache='none' to support migration, otherwise there is the chance of a hang and/or failure to pivot. However we prefer the default of cache=writethrough when operating in production. Our practice is to 'shutdown' the VM completely make the change with virsh edit, then restart. Then we have to repeat the process to revert back once we migrate. Is it possible to change that function on the fly and avoid the shutdown/start process? Note: we have had inconsistent results with virsh edit first and then a reboot. A complete shutdown seems to be necessary, but I am hoping there is some other procedure available. Our images are all qcow2 (1.1) Sincerely, W Kern PixelGate Networks

2 1

[libvirt-users] 回复： [libvirt] xml config nested
by 放牛班的春天 10 Jan '17

10 Jan '17

thanks your mail，now below： If the KVM supports nested VMX and QEMU starts with the arguments -enable-kvm and -cpu ..., + vmx, then the LOCK bit of the guest MSR_IA32_FEATURE_CONTROL and Enable VMX out of SMX operation bit will be set. Well, I should add -enable-kvm parameters of the document which xml position, Exactly speaking, how to amend xml configuration file? ------------------ 原始邮件 ------------------ 发件人: "Daniel P. Berrange";<berrange(a)redhat.com>; 发送时间: 2017年1月5日(星期四) 晚上8:53 收件人: "放牛班的春天"<446844717(a)qq.com>; 抄送: "libvirt-users"<libvirt-users(a)redhat.com>; 主题: Re: [libvirt] xml config nested NB, in future please don't CC all possible mailing lists at once. Just pick the most appropriate mailing list for your question. I've dropped libvirt-list & libvirt-announce from the CC, since this is a question most suited for libvirt-users. On Thu, Jan 05, 2017 at 11:44:29AM +0800, 放牛班的春天 wrote: > How to configure libvirt, so qemu-kvm support nested virtualization, virtual machine installed operating system is windows_7_ultimate_sp1_x64_dvd_618537.iso, configure libvirt xml file is as follows: > > <cpu mode='custom' match='exact'> > <model fallback='allow'>core2duo</model> > <feature policy='require' name='vmx'/> > </cpu> > > > or > > <cpu mode='host-model'> > <model fallback='allow'/> > </cpu> > > or > <cpu mode='host-passthrough'> > <topology sockets='2' cores='2' threads='2'/> > </cpu> Yes, that's the key guest configuration step - exposing the 'vmx' feature to the guest. In addition to that though, you need to make sure the kvm-intel kernel module in your host has the "nested=1" parameter set. eg in /etc/modprobe.d/kvm.conf you want options kvm_intel nested=1 If loaded correctly you should see # cat /sys/module/kvm_intel/parameters/nested Y if it says "N", then nested VMX will be disabled. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://entangle-photo.org -o- http://search.cpan.org/~danberr/ :|

1 0

[libvirt-users] xml config nested
by 放牛班的春天 10 Jan '17

10 Jan '17

How to configure libvirt, so qemu-kvm support nested virtualization, virtual machine installed operating system is windows_7_ultimate_sp1_x64_dvd_618537.iso, configure libvirt xml file is as follows: <cpu mode='custom' match='exact'> <model fallback='allow'>core2duo</model> <feature policy='require' name='vmx'/> </cpu> or <cpu mode='host-model'> <model fallback='allow'/> </cpu> or <cpu mode='host-passthrough'> <topology sockets='2' cores='2' threads='2'/> </cpu> We install libvirt in centos 7, qemu-kvm version is: [Root @ localhost libexec] # ./qemu-kvm --version QEMU emulator version 2.6.0 (qemu-kvm-ev-2.6.0-27.1.el7), Copyright (c) 2003-2008 Fabrice Bellard Above is the basic environment and libvirt configuration, but in this environment to install the internal implementation of the following windows7 virtual machine kernel interface calls are as follows: if (!FeatrueControlMsr.fields.enable_vmxon) { MyWriteFile(FileHanle, "Bios设置没有开启虚拟化\n", strlen("Bios设置没有开启虚拟化\n"), &ReturnLen); MyCloseFile(FileHanle); KdPrint(("Bios设置没有开启虚拟化")); return FALSE; } in conclusion: Through this kernel interface to determine the result is: Bios settings do not turn on virtualization How to solve this problem? Hope to get your help. thank you very much。

2 2

[libvirt-users] Regarding Migration Code
by Anubhav Guleria 09 Jan '17

09 Jan '17

Greetings, I was trying to understand the flow of Migration Code in libvirt and have few doubts: 1) libvirt talks to QEMU/KVM guests via QEMU API. So overall, in order to manage QEMU/KVM guests I can either use libvirt (or tools based on libvirt like virsh) or QEMU monitor. Is it so? 2) Since libvirt is Hypervisor neutral so actual migration algorithm(precopy or postcopy) is present in the hypervisor ,i.e. in case of QEMU it should be present in QEMU code base . I was going through the code starting from libvirt-domain api and not able to follow after virDomainMigrateVersion3Full(.....) . Kindly help. Thanks in advance. And sorry if that was too basic. Anubhav

2 2

[libvirt-users] libvirtError: block copy still active: disk not ready for pivot yet
by Ala Hino 05 Jan '17

05 Jan '17

Hi guys, When performing live merge, in few cases, we see the following exception: Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/vdsm/utils.py", line 736, in wrapper return f(*a, **kw) File "/usr/share/vdsm/virt/vm.py", line 5278, in run self.tryPivot() File "/usr/share/vdsm/virt/vm.py", line 5247, in tryPivot ret = self.vm._dom.blockJobAbort(self.drive.name, flags) File "/usr/share/vdsm/virt/virdomain.py", line 68, in f ret = attr(*args, **kwargs) File "/usr/lib/python2.7/site-packages/vdsm/libvirtconnection.py", line 124, in wrapper ret = f(*args, **kwargs) File "/usr/lib/python2.7/site-packages/vdsm/utils.py", line 1313, in wrapper return func(inst, *args, **kwargs) File "/usr/lib64/python2.7/site-packages/libvirt.py", line 733, in blockJobAbort if ret == -1: raise libvirtError ('virDomainBlockJobAbort() failed', dom=self) libvirtError: block copy still active: disk 'vdb' not ready for pivot yet That exception observed in following BZs: https://bugzilla.redhat.com/1376580 https://bugzilla.redhat.com/1397122 I am trying to understand what this exception indicates in order to handle it appropriately when thrown by libvirt. Thanks, Ala

2 3

[libvirt-users] how to configure seabios with libvirt
by dw 05 Jan '17

05 Jan '17

Hi: Anyone knows how to configure seabios with libvirt? Or Can I configure seabios in XML file? I know that could configre seabios with "qemu-system-x86_64 -fw_cfg",but how to with libvirt? thanks! dw

2 1