- Users - libvirt List Archives

Cannot passthrough usb with bus number and device number
by YaoHua Wu 08 Dec '20

08 Dec '20

Sometimes, qemu cannot find correct usb device number Reproduce steps: 1、lsusb Bus 001 Device 011: ID 090c:1000 1、prepare usb device attach xml <hostdev managed="yes" mode="subsystem" type="usb"> <source> <vendor id="0x090c" /> <product id="0x1000" /> <address bus='001' device='011'/> </source><address bus="1" port="1" type="usb" /> </hostdev> 2、virsh attach-device vm usb.xml error: internal error: Did not find USB device 090c:1000 bus:1 device 9 Question: why the device number returned in qemu's error report is not what I wrote in xml， what I mean is that if my dev number is really wrong, the returned error should also be the number I wrote， commonly when using dongle Look forward to your reply yaohua.wu

2 3

unable to migrate when TLS is used
by Vjaceslavs Klimovs 07 Dec '20

07 Dec '20

With libvirt 6.9.0, qemu 5.1.0, and following configurations: libvirt: key_file = "/etc/ssl/libvirt/server.lan.key" cert_file = "/etc/ssl/libvirt/server.lan.crt" ca_file = "/etc/ssl/libvirt/ca.crt" log_filters="3:remote 4:event 3:util.json 3:rpc 1:*" log_outputs="1:file:/var/log/libvirt/libvirtd.log" qemu: default_tls_x509_cert_dir = "/etc/ssl/qemu" default_tls_x509_verify = 1 migration with tls: virsh # migrate vm1 qemu+tls://server2.lan/system --persistent --undefinesource --copy-storage-all --verbose --tls never succeeds. Progress stops typically at high progress amounts (95%-98%), and network traffic drastically drops as well (from 1 gbps+ to nothing). domjobinfo progress also stops. Without --tls migrations succeed without issues without any other changes to hosts or configurations. Logs of failed migration: Source: https://drive.google.com/file/d/1d0dJumicW0TUdG1osNxNnWWiYpfAaIb_/view?usp=… Destination: https://drive.google.com/file/d/1d0dJumicW0TUdG1osNxNnWWiYpfAaIb_/view?usp=… Same exact hosts, successful migration logs (without --tls): Source: https://drive.google.com/file/d/1d0dJumicW0TUdG1osNxNnWWiYpfAaIb_/view?usp=… Destination: https://drive.google.com/file/d/1EWkCkSBhj76T05k86QjdL-6icJyruK5-/view?usp=… Should I report this as a bug or is there an issue with my configuration?

2 3

Nfs mount 'gone' after vm's start
by Marc Roos 06 Dec '20

06 Dec '20

I have had a nfs-ganesha mount provision install images for vm's, and recently I have the mount stall after starting a vm. (even when it does not access an iso image on the nfs mount) I have the host ip on a macvtap of the connected interface. The vm is also having an macvtap on the same interface to allow host-guest communication. I am mounting the nfs with these options. nfs4 ro,vers=4,proto=tcp,soft,nodev,_netdev,noatime,timeo=30,x-systemd.device -timeout=30,x-systemd.mount-timeout=30,x-systemd.automount,x-systemd.idl e-timeout=30 0 0 The problem is fixed by doing a simple 'umount -l' Anyone seen something similar? I suspect this could be related by the macvtap being created, but I cannot remember seeing this mount fail before.

1 0

apparmor DENIED on domain shutdown
by Francesc Guasch 05 Dec '20

05 Dec '20

Hi. I upgraded one of my servers to Ubuntu 20.04. Since then domains won't shutdown. They are in the "in shutdown" state. I googled around and I found it is probably because of apparmor. I see this message in the logs: kernel: [740222.848210] audit: type=1400 audit(1606983397.013:338): apparmor="DENIED" operation="signal" profile="libvirt-a2c1456f-3371-49eb-9fa4-f8576ca4e878" pid=2375 comm="libvirtd" requested_mask="receive" denied_mask="receive" signal=term peer="libvirtd" I tried to run this but it does nothing: aa-complain /usr/sbin/libvirtd The only solution I found was to run aa-disable to the machine id so it can be shut down. I found some similar issues but I can't figure out what I have to tweak in the apparmor files. If you can point me to some doc may be I skipped I will or any solution I would appreciate it. This is - Ubuntu 20.04 - libvirt 6.0.0-0 - qemu-kvm 1:4.2-3 Thank you very much.

2 3

qemu guest agent(QGA) in guest OS
by Henry lol 02 Dec '20

02 Dec '20

Hello, AFAIK, QGA is automatically installed and run in guest OS after installing OS from iso file whereas it's not installed in case of installing OS on the host even with same iso and configuration. So, I am wondering how it works differently depending on the guest or host. Does the OS installer selectively install packages? thanks.

1 0

How to exit console in L2 vm?
by Yalan Zhang 01 Dec '20

01 Dec '20

Hi, I have a question about nested virtualization. The scenario is as below: 1. Prepare the nested environment, start L2 guest. 2. On the host, connect the L1 vm console, then on L1 guest, connect the L2 guest console: (host)# virsh console L1_vm Connected to domain L1_vm Escape character is ^] (Ctrl + ]) ... (L1 vm)# virsh console L2_vm Connected to domain L2_vm Escape character is ^] (Ctrl + ]) ... (L2 vm)# <=== press " ^] " to exit the console, it return to the **host** (host)# Is it expected that "^]" in L2 guest will exit thoroughly to the host, not the L1 guest? Thank you! ------- Best Regards, Yalan Zhang IRC: yalzhang

4 4

qemu no-shutdown behavior in libvirt
by Luna Xu Yamada 30 Nov '20

30 Nov '20

2 2

Debugging virtiofs
by Popog 26 Nov '20

26 Nov '20

I can't seem to start a VM with virtiofs. Everything is running as root (unraid) with a pretty simple setup: <filesystem type='mount' accessmode='passthrough'> <driver type='virtiofs'/> <binary path='/usr/libexec/virtiofsd' /> <source dir='/mnt/user/shared'/> <target dir='shared_mount'/> </filesystem> Even with debug logging at max, all I get in the main log is: qemu-system-x86_64: -device vhost-user-fs-pci,chardev=chr-vu-fs0,tag=shared_mount,bus=pci.5,addr=0x0: Failed to write msg. Wrote -1 instead of 12. qemu-system-x86_64: -device vhost-user-fs-pci,chardev=chr-vu-fs0,tag=shared_mount,bus=pci.5,addr=0x0: vhost_dev_init failed: Operation not permitted shutting down, reason=failed And in the fsd log: [ID: 00029648] virtio_session_mount: Waiting for vhost-user socket connection... [ID: 00029648] virtio_session_mount: Received vhost-user socket connection [ID: 00000001] pivot_root(., .): Invalid argument Nothing else in any other logs (e.g. dmesg). What do these errors mean? How do I troubleshoot this?

1 0

image works in native but not in vm when cpu mode='host-passthrough' is set
by daggs 24 Nov '20

24 Nov '20

Greetings. I have an image I've created with a bunch of chost flags which works on my machine when it comes to native boot. if I take that same image into a vm managed via libvirt, I get kernel panic. I'd assume that something is missing from my vm config, question is what and what I can do about it? here is the flags part of lscpu in native and vm: https://dpaste.com/3TR8QJ5G8 and the vm's xml: https://dpaste.com/984JX4LUQ if I build the image with the default flags (march=x86-64), the vm boots well. my new chost is -O2 -pipe -march=skylake -mabm -mno-adx -mno-avx -mno-avx2 -mno-bmi -mno-bmi2 -mno-f16c -mno-fma -mno-xsave -mno-xsavec -mno-xsaveopt -mno-xsaves -mno-sgx the cpu is Intel(R) Pentium(R) CPU G4560 @ 3.50GHz thoughts? Thanks, Dagg.

3 5

Openshift Kubevirt lab?
by Luis Gonzalez 18 Nov '20

18 Nov '20

Is there an RHPDS lab for Kubevirt? -- Luis Gonzalez Consultant, Container Infrastructure Red Hat <https://www.redhat.com/> lgonzal(a)redhat.com M: 404.784.7166 IM: luisgonzalez GPG: 2800156E8386EA133D4F10AC3381C1457F69FF84 <https://www.redhat.com/>

2 1