December 2017 - Users - libvirt List Archives

[libvirt-users] libvirt-lxc nested containers
by ooth suresh 14 Dec '17

14 Dec '17

Hi, I am using libvirt to launch lxc containers. I would like to know if I can spawn nested container using libvirt-lxc. Is it supported ? Can some please shed some light on this ? Any pointers will be very helpful. Thanks Suresh S ‌

1 0

[libvirt-users] Inter Shared Memory ( Ivshmem ) : Cannot Bind
by Rogue S.T 13 Dec '17

13 Dec '17

Hello, friends. I encounter a problem when i use ivshmem with my guest, my ivshmem server is not start, and output a error : Example code, do not use in production ,cannot bind. Detail distribution: Today, I know ivshmem from a topic "QEMU version 2.10.93 User Documentation", Ivshmem: it can create a shared memory that guest device use, so we can use this memory to do something ourselves. First, I create host memory backend: -object memory-backend-file,size=1M,share,mem-path=/dev/shm/ivshmem,id=hostmem Twice, Bind a shared memory device to my guest: -device ivshmem-plain,memdev=hostmem Then I boot ivshmem servi ivshmem-server -p /var/run/ivshmem-server.pid -S /tmp/ivshmem_socket -M ivshmem -m /dev/shm -l 1M -n 2 Next time i get this error: *** Example code, do not use in production *** cannot bind And i search the problem on google, but nothing to get, hope you can help me solve it. following message is my host environment: host: ubuntu 1604 qemu: version 2.10.90 (v2.11.0-rc0-dirty) libvirt: libvirtd (libvirt) 3.9.0 guest: window 7 professor Refer: https://qemu.weilnetz.de/doc/qemu-doc.html#Inter_002dVM-Shared-Memory-device (topic qemu documention) https://helpmanual.io/help/ivshmem-server/ ( how to use ivshmem-server )

2 1

[libvirt-users] libvirt authentication
by Anastasiya Ruzhanskaya 11 Dec '17

11 Dec '17

Hello, I am currently trying to install certificates for tls. By this time I have got some questions: 1) Is documentation in the web docs up-to-date regarding tls server, client, ca certificates? (actually I have some problems, but maybe this is due to smth has changed in certtools and was not updated in docs, perhaps some more fields are needed now). 2) there is a error in virt-pki-validate : when determining whether host name matches CN name you assume, that CN is standing in the end. So I get an error: The server certificate does not seem to match the host name hostname: "barni" Server certificate CN: "barni,O=libvirt.org" CN field is standing at the beginning of line. Seems that these are the appropriate lines in code: 258 S_HOST=`"$CERTOOL" -i --infile "$LIBVIRT/servercert.pem" | grep Subject: | sed 's+.*CN=$[a-zA-Z\. _-]*$+\1+'` 259 if test "$S_HOST" != "`hostname -s`" && test "$S_HOST" != "`hostname`"

1 1

[libvirt-users] libvirt error messages
by Anastasiya Ruzhanskaya 11 Dec '17

11 Dec '17

Hello, I am experimenting with libvirt for my diploma. I set up an environment, where I can stop, resend, generate messages for libvirt. I am capable also standing between client and server and analyze the number of current procedure and decide what to do next ( I mean rpc procedure). So, for example I want to always drop creating snapshots and send an error to client with my custom message (NOT custom error type, but exactly custom message, I don't want to modify any libvirt code on client and adding my errors). I am operating with bare tcp traffic, so I can easily read everything in the middle. So, I noticed one error for libvirt snapshots: 'internal error: unexpected domain snapshot snapshot1 already exists' and started sending it every time I see procedure : REMOTE_PROC_DOMAIN_SNAPSHOT_CREATE_XML I thought, that this is possible to insert custom message here, but this is not true. When I change the length of the message on one symbol - it is still correct, on two -also , on three symbols - I get an error that the size of the payload is incorrect. Some other errors when I insert a message of random length. What is strange, that if I insert 'xxxxxxxxxxxxx...' message , which has the same length as 'internal error: unexpected domain snapshot snapshot1 already exists' - everything is parsed fine. So, how does this stuff work? Are there anywhere predefined lengths for error messages and checks against them, or are there checks against exact error strings? How all this stuff is regulated? I know now, that I can pad my message with ...say, space symbols and it will be ok, but I want to know how error parsing works in all cases. I have looked through files, where an overall structure and logic of errors is described but didn't find any helpful information. And I don't a good way to debug libvirt to see, where this happens.

2 1

[libvirt-users] public IP for VM
by scar 08 Dec '17

08 Dec '17

I can't seem to figure out how to setup a public IP for the VM so we could use it as a second public webserver. The host machine runs our primary webserver with CentOS 6.9 and one public ipv4 address configured on physical interface eth0. Now i have configured a new bridge interface br0: DEVICE=br0 TYPE=Bridge BOOTPROTO=static The IP settings are for a second public ipv4 address in the same network. Now i setup the VM to use br0 for the NIC and set the same IP settings inside the VM... but it doesn't work. The VM can't resolve any hosts, and if i try to ssh to the second public IP, it just sends me to the host machine instead of the VM.

5 4

[libvirt-users] public IP for VM
by Chandler 08 Dec '17

08 Dec '17

I can't seem to figure out how to setup a public IP for the VM so we could use it as a second public webserver. The host machine runs our primary webserver with CentOS 6.9 and one public ipv4 address configured on physical interface eth0. Now i have configured a new bridge interface br0: DEVICE=br0 TYPE=Bridge BOOTPROTO=static The IP settings are for a second public ipv4 address in the same network. Now i setup the VM to use br0 for the NIC and set the same IP settings inside the VM... but it doesn't work. The VM can't resolve any hosts, and if i try to ssh to the second public IP, it just sends me to the host machine instead of the VM. I read the Bridged Networking section but it seems to want to set br0 to the public IP and not use eth0. I want to use two public IP's over the same shared phys interface, eth0 with our first IP and br0 with the second IP for the VM... or another similar configuration that will have the same effect. Thanks

2 1

[libvirt-users] How to use virtio-gpu to output
by Rogue S.T 07 Dec '17

07 Dec '17

Hello, lads, owing to i study video card on my guest recently. So i have a notion that whether we can get video with virtio-vga or not. It's seem that we are using a graphics card truth. Maybe we can finish? And if this is not likelihood? Tell me why please. Thanks following is my host system environment: host: ubuntu 1604 qemu: 2.11 libvirt: 3.9 kernel: 4.14 guest: window 7 professor

1 0

[libvirt-users] problem when use tls to connect libvirt
by Yalan Zhang 06 Dec '17

06 Dec '17

Hi guys, I met a problem when I use tls to connect libvirt. When I set the CN in client.info, server.info as hostname(FDQN), the tls check will fail with ip; and vice versa, when set CN as ip address, the tls check will fail with hostname. Only use what we set in can succeed. If this is expected? or I there was some issue in my env. or setup steps? 1. set tls env with hostname, then it will fail to check with ip # virsh -c qemu+tls://192.168.122.4/system 2017-12-06 13:24:52.346+0000: 3954: info : libvirt version: x.x.x, package: 4.el7 (Red Hat, Inc. <http://bugzilla.redhat.com/bugzilla>, 2017-11-30-07:57:27, x.x.x.redhat.com) 2017-12-06 13:24:52.346+0000: 3954: info : hostname: work.englab.cn 2017-12-06 13:24:52.346+0000: 3954: warning : virNetTLSContextCheckCertificate:1125 : Certificate check failed Certificate [session] owner does not match the hostname 192.168.122.4 error: failed to connect to the hypervisor error: authentication failed: Failed to verify peer's certificate 2. use the hostname as what we set can succeed. # virsh -c qemu+tls://test.englab.cn/system Welcome to virsh, the virtualization interactive terminal. Type: 'help' for help with commands 'quit' to quit virsh # # ping test.englab.cn PING test.englab.cn (192.168.122.4) 56(84) bytes of data. 64 bytes from test.englab.cn (192.168.122.4): icmp_seq=1 ttl=64 time=0.235 ms 64 bytes from test.englab.cn (192.168.122.4): icmp_seq=2 ttl=64 time=0.204 ms ... ------- Best Regards, Yalan Zhang

2 1

[libvirt-users] Libvirt checkout graphics card
by Rogue S.T 06 Dec '17

06 Dec '17

Hello, everybody!My guest has two graphics card when it is running, one is GTX960, and other is video card that is created by qemu.But my guest use video card default, and my screen is not light. I try checkout graphics card with vnc or spice, But it is suspend.Could you help me? And may i checkout card in my host system with command? Because my connection with vnc or spice is not stable.This is my vga card message: Bus 0, device 7, function 0: VGA controller: PCI device 1234:1111 BAR0: 32 bit prefetchable memory at 0xf2000000 [0xf2ffffff]. BAR2: 32 bit memory at 0xfe09e000 [0xfe09efff]. BAR6: 32 bit memory at 0xffffffffffffffff [0x0000fffe]. id "video0" Bus 0, device 8, function 0: VGA controller: PCI device 10de:1401 IRQ 11. BAR0: 32 bit memory at 0xfd000000 [0xfdffffff]. BAR1: 64 bit prefetchable memory at 0xe0000000 [0xefffffff]. BAR3: 64 bit prefetchable memory at 0xf0000000 [0xf1ffffff]. BAR5: I/O at 0xc000 [0xc07f]. BAR6: 32 bit memory at 0xffffffffffffffff [0x0003fffe]. id "" And following message is my envireoment: host: ubuntu 1604 guest: window 7 professor qemu: 2.11.0-rc0 seabiois: 2.10.7 (qemu set) libvirt: 3.9

1 0

[libvirt-users] Guest Reboot After Plug HDMI
by Lying 05 Dec '17

05 Dec '17

Hello, Everybody. I encounter a problem when i using my guest. I booting my guest without HDMI primarily, Then i add it, but my guest is be rebooted. To know what cause it, i do it again, especially i check my guest is running normal or not with "ping" before plugging, that's well. What i was suffer? can you help me solve it? Following message is detailed!!! Guest: window 7 profession Linux: Ubuntu 1604 Qemu:1 v2.11.0-rc0 Libvirt: 3.9 method of booting: virsh start myGuest1 method of plug HDMI: virsh qemu-monitor-command myGuest1 --hmp 'device_add driver=vfio-pci,host=01:00.0,xvga=on,multifunction=yes' phenomenon: after plugging, my Guest is failed to check with "ping", for a while, my display is light and load windows. expect: my guest is running when i plug HDMI, it seem that hot add a device.

3 2