- Users - libvirt List Archives

[libvirt-users] migrated RHEL7/CentsOS7 VMs fail to boot
by Daniel Pocock 03 Apr '17

03 Apr '17

I migrated a CentOS7 VM (the fedrtc.org site) from an XCP environment to a libvirt/KVM environment. This involved using qemu-img to convert the image from VHD to qcow2 and then using virt-install --import to define the VM in libvirt. Three problems occurred during boot: a) on the first boot, the BIOS screen and grub screen don't appear at all, the screen is blank for a couple of seconds and then the message about loading a kernel appears. Hard reset and on the second and subsequent attempts I see the grub screen and have the opportunity to interact with it. This actually happened with many of my VMs, not just the CentOS7 VM. b) with console=hvc0 in the grub config, the kernel refused to boot, no error appeared on the screen. I was able to remove that easily enough by pressing "e" in grub. Rather than halting, should the kernel fall back to VGA perhaps when the console= argument is not valid? Or could KVM emulate the Xen console device to make migrations easier? c) after that, the boot proceeds up to about the point where I see systemd messages about starting basic system. Then it sits there for a few minutes and then these messages appear: warning: dracut-initqueue timeout - starting timeout scripts and after that I see "A start job is running for dev-mapp...oot.device (X min Ys/ no limit) Rebooting and choosing the "rescue" image from the bottom of the grub menu got me past that, I was able to log in and then I ran: dracut --kver 3.10.0-514.10.2.el7.x86_64 --force and on the next attempt it was able to boot successfully. The block device for the root FS is an LVM volume (so the name should have been constant) and the block device for the /boot filesystem listed in /etc/fstab was mounted by UUID (the block device name itself changed from xvda1 (XCP) to vda1 (KVM)). All my Debian VMs were able to cope with this device name changing. The CentOS7 system was originally installed using default settings for just about everything. Why does dracut need to be re-run in this situation? Should a bug report be filed about this?

2 2

[libvirt-users] different bios file per guest
by Anton Gorlov 31 Mar '17

31 Mar '17

Can I set different bios file per guest? For some guest need UEFI. By default using bios-256k.bin

2 1

[libvirt-users] libvirt + lxc + cgroups mounting
by Pyles, Andy 22 Mar '17

22 Mar '17

All, Within: https://libvirt.org/drvlxc.html#cgroups it mentions that : “/sys/fs/cgroup/NNNN the host cgroups controllers bind-mounted to only expose the sub-tree associated with the container” In my environment, the host contains: $ls –l /sys/fs/cgroup/ cpu -> cpu,acct cpu,cpuacct cpuacct -> cpu,cpuacct cpuset debug devices freezer memory net_cls systemd - but on my LXC container, I have only: $ls –l /sys/fs/cgroup memory -- How do I add additional host cgroup controllers to the set which is bind-mounted within the container? Thanks, Andy Pyles

1 0

[libvirt-users] libvirtd firewalld support
by Leon Goldberg 22 Mar '17

22 Mar '17

Hey, I've been wondering about the extent libvirt makes use of firewalld. I'm looking to use firewalld exclusively and wonder about the ports libvirt takes care of for me via firewalld (e.g. will console/vnc ports, qemu migration ports, etc, will be handled for me?) I have only done a little bit of testing, but it seems like I can't connect to a vm via vnc without explicitly opening ports beforehand. Thanks, Leon

3 2

[libvirt-users] Memory statistics for a domain
by Szczerbik, PrzemyslawX 22 Mar '17

22 Mar '17

Hi, I'm wondering whether there is any difference between VIR_DOMAIN_MEMORY_STAT_ACTUAL_BALLOON and 'memory' member from virDomainInfo structure. Based on the documentation those values should correspond to current balloon and memory used by the domain respectively. Is it always the case that those values will be equal? Is there a reason why maximum memory allowed ('maxMem' from virDomainInfo structure) cannot be retrieved from virDomainMemoryStats API? It seems that it's required to call 2 different APIs to get all memory metrics for a domain. Regards, Przemek -------------------------------------------------------------- Intel Research and Development Ireland Limited Registered in Ireland Registered Office: Collinstown Industrial Park, Leixlip, County Kildare Registered Number: 308263 This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

1 0

[libvirt-users] Running virConnectGetCapabilities / virsh capabilities without a daemon or KVM
by Richard W.M. Jones 22 Mar '17

22 Mar '17

For virt-p2v we want to get the host CPU model, topology etc, and the best way we found to do that was to pull it from libvirt capabilities[1]. Since libvirt is already parsing the host CPU information, and because /proc/cpuinfo is such a stupid format to parse, this is very convenient. Unfortunately it doesn't work unless libvirtd _and_ KVM are both installed. Without libvirtd installed: # virsh capabilities error: failed to connect to the hypervisor error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': No such file or directory With libvirt-daemon installed, but not QEMU: # virsh capabilities error: failed to connect to the hypervisor error: no connection driver available for <null> I feel I should be able to fiddle with the connection URI in some way to make this possible, at least in the second case. However there's not a concept of "no hypervisor", and I couldn't find any other way to make it work. Using "test:///default" seems tempting, but it substitutes some imaginary host (i686) so that's no good. Installing KVM isn't really an option for us since we are trying to build a minimal bootable ISO. Just installing qemu-system-x86_64 pulls in half the world. (Even installing libvirtd is difficult, and we'd prefer to avoid it). Any ideas? Rich. [1] https://github.com/libguestfs/libguestfs/commit/963d6c3be7cd91c0373f67cfdd9… -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com Fedora Windows cross-compiler. Compile Windows programs, test, and build Windows installers. Over 100 libraries supported. http://fedoraproject.org/wiki/MinGW

3 3

[libvirt-users] libvirt-guests and time sync on autostart vms
by ıuoʎ 21 Mar '17

21 Mar '17

I have a host running a few guests all configured to autostart. I'm also using libvirt-guests to suspend the guests when the host restarts or shuts down When the host comes back up I see a bunch of log messages - Resuming guest vm01 already active Resuming guest vm02 already active I guess this is because libvirt autostarts the guests when it starts and before libvirt-guests has a chance to start them. I didn't think this is an issue until I noticed the time on the guests is totally out of sync. SYNC_TIME=1 is enabled in the conf file but this doesn't seem to work I can see from the code at libvirt-guests.sh that the time sync function will only run if the guest is started by libvirt-guests ans will be skipped for guests that are already running So I'm not sure if this is a bug in libvirt-guests.sh or if I'm doing something wrong by using both libvirt-guests and autostart. I assume libvirt-guests will always resume guests on proper reboot but I'm think if there is any unexpected failure where guests weren't properly suspended before boot I still want them to come backup when the server boots Regards, Yoni

3 4

Re: [libvirt-users] CPU Pinning Help
by arrakis＠tsiom.com 21 Mar '17

21 Mar '17

1 0

[libvirt-users] [RFC] per-device metadata
by Francesco Romani 21 Mar '17

21 Mar '17

Hi, Currently libvirt supports metadata in the domain XML. This is very convenient for data related to the VM, but it is a little awkward for devices. Let's pretend I want to have extradata (say, a specific port for a virtual switch) related to a device (say, a NIC). Nowadays I can store that data in the metadata section, but I need some kind of mapping to correlate this piece of information to the specific device. I can use the device alias, but this is not available when the device is created. This is also more complex when doing hotplug/hotunplug, because I need to do update device and update metadata; if either fails, the entire operation must be considered failed. It would be nice to be able to attach metadata to the device, and this is what I'm asking for/proposing in this mail. Would it be possible in a future libvirt release? If this is not possible, what's the best way to do the aforementioned mapping? If it's alias (or device address), how can I be sure than I'm addressing (no pun intended) the right device when I don't have them? (e.g. hotplug new device, or just first time VM created)? Thanks, -- Francesco Romani Red Hat Engineering Virtualization R & D IRC: fromani

3 9

[libvirt-users] Get PID of a domain's QEMU instance from its domain ID
by Thibaut SAUTEREAU 21 Mar '17

21 Mar '17

Hello, I cannot find a way to retrieve PIDs of QEMU instances from libvirt domains' IDs (I'm using libvirt C API). I recognize it sounds like a bad idea doing so (and I know PIDs are explicitly made not available, as I gathered from the source code and on your IRC channel) but I need that to use the perf_event_open syscall in order to gather statistics on my QEMU/KVM guests. I also know libvirt now supports some perf events but only a few and I need more. I could submit patches to add them and I will definitely consider that but in the meantime... What would be the best way to get those PIDs? I tried using the XML file but PID is hidden in there too. I took a look at QEMU Machine Protocol. Now I'm going to walk /proc and match on guests names but it is not that elegant. Any ideas? Thanks, Thibaut S.

3 3