- Users - libvirt List Archives

[libvirt-users] Inconsistent behavior between x86_64 and ppc64 when creating guests with NUMA node placement
by Michael Turek 15 Sep '14

15 Sep '14

Hello all, I was recently trying out NUMA placement for my guests on both x86_64 and ppc64 machines. When booting a guest on the x86_64 machine, the following specs were valid (obviously, just notable excepts from the xml): <memory unit='KiB'>8388608</memory> <currentMemory unit='KiB'>8388608</currentMemory> <vcpu placement='static'>4</vcpu> ... <cpu> <topology sockets='4' cores='1' threads='1'/> <numa> <cell cpus='0-2' memory='6144'/> <cell cpus='3' memory='2048'/> </numa> </cpu> However, on ppc64 this causes the following error: error: Failed to create domain from sample_guest.xml error: internal error: early end of file from monitor: possible problem: 2014-09-11T18:44:25.502140Z qemu-system-ppc64: total memory for NUMA nodes (8388608) should equal RAM size (200000000) The 200000000 is actually 8192 MB in bytes and hexidecimal. This is apparently just an issue with the error message. The following specs work on ppc64: <cpu> <topology sockets='4' cores='1' threads='1'/> <numa> <cell cpus='0-2' memory='6291456'/> <cell cpus='3' memory='2097152'/> </numa> </cpu> Note that the memory for each cell is 6144*1024 and 2048*1024 respectively. The issue is that the memory size for each NUMA cell should be specified in KiB, not MB (http://libvirt.org/formatdomain.html#resPartition "|memory| specifies the node memory in kibibyte"). In short, it seems that specifying NUMA cell memory in MB works on x86_64 but not on ppc64. Does anyone have any insight to what's causing this, or if I'm misunderstanding something? Any help is appreciated, thank you! Regards, Michael Turek

2 1

[libvirt-users] ntpd in VM
by Adam King 12 Sep '14

12 Sep '14

----- Original Message ----- From: "Adam King" <kinga(a)sghs.org.uk> To: "Gary Hook" <garyrhook(a)gmail.com> Sent: Friday, September 12, 2014 1:39:38 PM Subject: Re: [libvirt-users] ntpd in VM ----- Original Message ----- From: "Gary Hook" <garyrhook(a)gmail.com> To: libvirt-users(a)redhat.com Sent: Friday, September 12, 2014 1:25:47 PM Subject: Re: [libvirt-users] ntpd in VM While I agree that running a time server in a VM is, at best, problematic, most of those nay-sayers have experience with VMware, xen and the like. Those aren't the only hypervisors out there, and the decision process should depend upon the hypervisor (to a significant degree), not just the idea of "can I do this in a VM?" Not all hypervisors are created equally. On Fri, Sep 12, 2014 at 7:14 AM, Pierre Schweitzer < pierre(a)reactos.org > wrote: Hi, It is still a bad idea. I invite you to read here for the reasons why: http://serverfault.com/a/106509/150152 Cheers, On 09/12/2014 12:51 PM, Mauricio Tavares wrote: > I was taught in kitty school that running a ntp server in a vm > was a bad idea. Is that still the case? > My experience is somewhat different. VMWare has always been problematic with this, especially Linux guests. I've run KVM based NTP servers inside vm's (2 as active:standby) and dozens of ntp clients inside vm's for over a year now and seen no adverse results. Admittedly I've not been checking the drift but I can vouch for the results. Regards Adam King <blockquote> </blockquote>

1 0

[libvirt-users] ntpd in VM
by Mauricio Tavares 12 Sep '14

12 Sep '14

I was taught in kitty school that running a ntp server in a vm was a bad idea. Is that still the case?

3 2

[libvirt-users] trying but failing - pass-through scsi/sata
by lejeczek 11 Sep '14

11 Sep '14

hi everybody I'd imagine this topic must have been covered, yet I've failed to find any really helpful and a solid howto/examples of how to pass-through a sata disks. currently I'm having this in my xml: <hostdev mode='subsystem' type='scsi' managed='no'> <source> <adapter name='scsi_host4'/> <address bus='0' target='0' unit='0'/> </source> <address type='drive' controller='4' bus='0' target='0' unit='0'/> </hostdev> I took it from http://libvirt.org/formatdomain.html#elementsHostDevSubsys yet a Linux guest (SL7) fails to find/detect any disks, if I dumpxml I get this: <device> <name>block_sdb_OCZ_VERTEX4_OCZ_352C77TZVQF1Y3SX</name> <path>/sys/devices/pci0000:00/0000:00:1f.2/ata5/host4/target4:0:0/4:0:0:0/block/sdb</path> <parent>scsi_4_0_0_0</parent> <capability type='storage'> <block>/dev/sdb</block> <bus>ata</bus> <drive_type>disk</drive_type> <model>OCZ-VERTEX4</model> <vendor>ATA</vendor> <serial>OCZ-VERTEX4_OCZ-352C77TZVQF1Y3SX</serial> <size>128035676160</size> <logical_block_size>512</logical_block_size> <num_blocks>250069680</num_blocks> </capability> </device> how do I pass-through sata drives? any help much appreciated, thanks!

1 0

[libvirt-users] QEMU disk migration not using entire connection throughput
by Joaquim Barrera 08 Sep '14

08 Sep '14

Hello everybody, I've been observing that, during the disk migration usign --copy-storage-inc, the data transfer hardly ever uses the entire available connection bandwidth (in my case 1 Gb/s), and the migration is at a non-constant speed. On the other hand, at the memory migration step, everything goes perfect. I am almost sure that this is not a nard disk throughput limitation, because a 'scp' command between the two same machines performs perfectly right. Is libvirt (or maybe QEMU?) doing some extra work that may cause slowlyness? Thanks!

1 0

[libvirt-users] USB redirection
by Fl＠sh 05 Sep '14

05 Sep '14

What is the difference between 'USB Host device assignment' from 'USB Redirected devices'? -- Fl@sh

2 1

[libvirt-users] Default lib/include dirs location in OS X
by Radek Simko 05 Sep '14

05 Sep '14

Hi, I have installed libvirt via Homebrew = compiled from source <https://github.com/Homebrew/homebrew/blob/master/Library/Formula/libvirt.rb> . Then I wanted to install *ruby-libvirt* gem install ruby-libvirt and end up with this error: /usr/local/Cellar/ruby/2.1.2_3/bin/ruby extconf.rb *** extconf.rb failed *** Could not create Makefile due to some reason, probably lack of necessary libraries and/or headers. Check the mkmf.log file for more details. You may need configuration options. Provided configuration options: --with-opt-dir --with-opt-include --without-opt-include=${opt-dir}/include --with-opt-lib --without-opt-lib=${opt-dir}/lib --with-make-prog --without-make-prog --srcdir=. --curdir --ruby=/usr/local/Cellar/ruby/2.1.2_3/bin/ruby --with-libvirt-include --without-libvirt-include --with-libvirt-lib --without-libvirt-lib --with-libvirt-config --without-libvirt-config --with-pkg-config --without-pkg-config extconf.rb:73:in `<main>': libvirt library not found in default locations (RuntimeError) extconf failed, exit code 1 so then I passed the locations of *include* & *lib* directories as the configuration options: gem install ruby-libvirt -- --with-libvirt-include=/usr/local/Cellar/libvirt/1.2.7/include --with-libvirt-lib=/usr/local/Cellar/libvirt/1.2.7/lib which *worked*, but it's not very clean and effective solution, especially if you have *ruby-libvirt* as a requirement in your Gemfile and you need to install it via Bundler. I reckon that a possible solution would be to change Homebrew formula, so that it either installs or symlinks both dirs into locations where *ruby-libvirt* will look for it, but *what's the default location actually*? -- *Radek Simko*

2 1

[libvirt-users] Unable to find security driver for label selinux
by Qiang Guan 05 Sep '14

05 Sep '14

Hi experts, I want to have a test on security driver for libvirt lxc on my debian system. What I do is as the following steps: 1) download the source code from git://libvirt.org/libvirt.git 2) compile and install with the source code as following: ./autogen.sh --system ./configure --with-selinux=yes --with-secdriver-selinux=yes make -j8 & make install root@debian:~/github/libvirt.git/tools# ./virsh --version=long Virsh command line tool of libvirt 1.2.8 See web site at http://libvirt.org/ Compiled with support for: Hypervisors: QEMU/KVM LXC UML OpenVZ VMWare VirtualBox Test Networking: Remote Network Bridging Interface udev Nwfilter VirtualPort Storage: Dir Filesystem SCSI Multipath LVM Miscellaneous: Daemon Nodedev SELinux Secrets Debug Modular 3) then I define a lxc vm with the seclabel : root@debian:~/images# vir dumpxml lxc <domain type='lxc'> <name>lxc</name> <uuid>b1b787a1-d20e-48bd-938b-16ba61d22405</uuid> <memory unit='KiB'>419404</memory> <currentMemory unit='KiB'>419404</currentMemory> <vcpu placement='static'>1</vcpu> <resource> <partition>/machine</partition> </resource> <os> <type arch='x86_64'>exe</type> <init>/sbin/init</init> <cmdline>console=tty0 console=ttyS0</cmdline> </os> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/local/libexec/libvirt_lxc</emulator> <filesystem type='mount' accessmode='passthrough'> <source dir='/tmp/rootfs'/> <target dir='/'/> </filesystem> <console type='pty'> <target type='lxc' port='0'/> </console> </devices> <seclabel type='dynamic' relabel='yes'/> </domain 4) When I start the vm, It output an error: root@debian:~/images# vir start lxc error: Failed to start domain lxc error: unsupported configuration: Unable to find security driver for label selinux What's the problem? Really appreciate for any help. -- ------------ Jackie Best Regards

2 1

[libvirt-users] virtio-serial port that listens tcp socket on host
by Vasiliy Tolstov 05 Sep '14

05 Sep '14

Hello. Is that possible to create virtio-serial port that on host listens tcp socket and how to avoid duplicate listened ports in case of migration? -- Vasiliy Tolstov, e-mail: v.tolstov(a)selfip.ru jabber: vase(a)selfip.ru

3 5

[libvirt-users] Fwd: Does virsh actually work?
by Gary Hook 04 Sep '14

04 Sep '14

On Wed, Sep 3, 2014 at 2:12 PM, Eric Blake <eblake(a)redhat.com> wrote: > On 09/03/2014 01:05 PM, Gary Hook wrote: > > > > > root@huey:/tmp# virsh migrate --live --p2p > > This command does not default to migrating storage,... > Oh, you mean like this? root@huey:/tmp# virsh migrate --live --p2p --copy-storage-all e7f75b9b-9ed4-4f7e-aa86-e481ab911d6f qemu+ssh://10.10.11.204/system --verbose error: Failed to open file '/mnt/store01/virt/e7f75b9b-9ed4-4f7e-aa86-e481ab911d6f.qcow2': No such file or directory Apologies; I should have used this one in the first place. Too many attempts in the command history.... I'm guessing I"m still missing something... > > > > (huey is 10.10.11.203, dewey is 10.10.11.204; DNS works, but I thought to > > eliminate that as a source of this problem.) As can be seen, the .qcow2 > > files exists on the source system, and does not exist on the target > system. > > ...therefore, this error message is expected if the storage is not > already present on the destination. > > Migration requires either that you have shared storage (both source and > destination can see THE SAME FILE, and if one side sees the file via > NFS, then both sides must see the file via NFS [you can't mix local and > NFS]). OR, you have to tell migration that you want to migrate storage > as well as machine state, by adding the --copy-storage-all flag to your > command line. > Understood; perfectly sensible. See above. > http://libvirt.org/migration.html should document useful information > such as this; would you like to submit a patch to help improve that web > page? The sources are in libvirt.git under docs/migration.html.in. > Yep, read that page (several times). At such point I have anything of substance to add, absolutely, yes, I very much want to contribute. I am just starting out building my skills on KVM and libvirt (background is with PHYP on Power) but fully intend to become conversant with the packages.

2 1