- Users - libvirt List Archives

[libvirt-users] libvirt live migrate problem
by lyz_pro 02 Apr '14

02 Apr '14

Hi The problem is after I successful live migrate a domain, the domain os become read-only. and when I use qemu-img to show the disk info at node which domain current running , the command return 'Permission denied'. BUT, I can use qemu-img show the disk info at the previous node. My Env: two node, all install libvirt + kvm . use glusterfs as shared storage. the libvirt/kvm version: libvirt-0.10.2-29.el6_5.5.x86_64 qemu-kvm-0.12.1.2-2.415.el6_5.6.x86_64 qemu-img-0.12.1.2-2.415.el6_5.6.x86_64 glusterfs-3.4.1-3.el6.x86_64 -------------- lyz_pro 2014-04-02

2 1

[libvirt-users] ebtables rules are not applied when using libvirt nwfilter
by Thinh Nguyen 02 Apr '14

02 Apr '14

Dear all, I configure my kvm vm like this: <interface type='bridge'> <mac address='52:54:00:dd:b2:c5'/> <source bridge='nw-vpc-1017'/> <target dev='if-57'/> <model type='virtio'/> <filterref filter='clean-traffic'> <parameter name='IP' value='10.0.0.1'/> </filterref> <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> </interface> But when i start it and show the ebtables rules, nothing is applied [root@kvmhost ~]# ebtables -L Bridge table: filter Bridge chain: INPUT, entries: 0, policy: ACCEPT Bridge chain: FORWARD, entries: 0, policy: ACCEPT Bridge chain: OUTPUT, entries: 0, policy: ACCEPT Bridge chain: libvirt_qemu_FORWARD, entries: 0, policy: DROP Please show me what wrong i've done. Thanks so much! -- Nguyen Thinh

2 1

[libvirt-users] Advice request for building a centralized vm disk provider server
by Pasquale Dir 02 Apr '14

02 Apr '14

Hello again, sorry for mailing in so short time but I have another question about live migration. >From redhat documentation I see that the storage for a VM which has to be migrated has to be shared throught some mechanism and I have chosen iscsi. Anyway I now would need a way to dinamically create a new virtual machine and so a virtual hard disk, note that this vm will have to be migrated too. I though about this solution: I would share an HUGE iscsi target, this target is a logical hard drive containing many disk images which would be the VM's hard drives. I can easily clone a vm by making a simple file copy of a vm disk. This would allow me to create a VM by specifying a path like <iscsi target path>/disk1.img. I have two problems: 1)how to get an unique <iscsi target path>? On my phisical host it is mounted as a random uuid in the /media folder. 2)will migration work with a <file> instead of a <device> as hard disk field? If you have any better solution, even involing the change from iscsi to another technology, you are free to propose it to me. Sorry for bothering, Pasquale

1 0

[libvirt-users] Async migration: possible?
by Pasquale Dir 02 Apr '14

02 Apr '14

I am seeing that migrate() function is blocking, isn't there a non blocking function with a callback associated to it, so to report events like end migration, migration progress (job domain info informations) and so on? Pasquale

2 2

[libvirt-users] Ask for help about bug 1058149
by Qiang Guan 01 Apr '14

01 Apr '14

On Mar 31, 2014 at 21:23PM, Martin Kletzander wrote: > On Thu, Mar 27, 2014 at 10:07:43AM +0800, Qiang Guan wrote: >> Hi Martin, >> >> Is there any plan to fix the libvirt bug 1058149 which I reproduced on >> the libvirt master branch? >> > > Hi, > > Sorry for such late response, I'm currently busy working on whole > different part of the codebase. I'll try to see what I can do, but > you can always try asking on the list. Is there some version you were > not able to reproduce it on? > > Martin I've succeeded reproducing this problem on v1.1.4-maint ,v1.2.1-maint, v1.2.2 and the master branch. It seems that the problem always exsisted for libvirt under my testing scene. -- ------------ Jackie Best Regards

1 0

[libvirt-users] Programmatically force shutdown a guest: possible?
by Pasquale Dir 01 Apr '14

01 Apr '14

I am looking at the shutdown method, but if the guest system is a desktop system, like for example ubuntu, it just has the effect to show a box prompting the user for a shutdown/reboot/ and such. I could enter the guest and change this default behaviour and it actually works..but I'd like for a way to send a shutdown command without doing so. Is it possible?

3 4

[libvirt-users] Connecting libvirt to manually compiled QEMU
by Asadullah Hussain 31 Mar '14

31 Mar '14

Hello I have manually compiled a customized qemu (1.4.0) which runs fine on its own (create VM etc) but I want to access this qemu through libvirt (virt-manager, virsh etc). But the libvirt driver only looks into "/usr/bin" for qemu binaries, how can I tell libvirt to connect to my qemu which is placed at "/home/user/qemu" directory. -- Asadullah Hussain

3 6

[libvirt-users] Specify a disk image file which is on an iscsi target
by Pasquale Dir 31 Mar '14

31 Mar '14

Hello, for my vm I'd like to specify a disk image which would be on an iscsi target. Currently, by using targets directly, I'd just specify the path in /dev/disk/by-path/ip-192.168.1.2:3260-iscsi-iqn.myiqn but how to specify a file which is actually INSIDE this path, ex. /dev/disk/by-path/ip-192.168.1.2:3260-iscsi-iqn.myiqn/myVmImage.img ? I saw that there is another path in /media/myHomeUser/anId but do this id changes each time?

1 0

[libvirt-users] virsh attach-disk live guest problem
by john fisher 29 Mar '14

29 Mar '14

Hypervisor= ubuntu 12.04.2 guest = ubuntu 12.04.4 used virt-manager to create a new storage file in qcow2 format on hypervisor and then attach it to the guest. In virt-man I can see the disk is attached. trying virsh attach-disk I get a failure saying disk already attached virsh dump xml shows the disk at vdb as it should be but on the guest, fdisk -l shows no vdb and its not in /proc I found one reference to apparmor saying add /proc rw to hypervisor's apparmor libvirt conf But I see some release notes that might be about avoiding this ( sounds insecure) so... Do I have to stop and restart the guest, change apparmor or what? -- John Fisher

1 1

[libvirt-users] Using qemu+ssh on openSUSE 13.1/Tumbleweed
by Johannes Kastl 29 Mar '14

29 Mar '14

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi everyone, I am trying to get libvirt with qemu-kvm to work on my machines running openSUSE 13.1 / Tumbleweed. My question is in regard to using qemu+ssh, which would be my preference, as I already have a working ssh-key authentication with SSH-Agent. I set the permissions of the manage-socket to 0770, added my user to the libvirt group and set auth_unix_rw = "none" in /etc/libvirt/libvirtd.conf. I am now able to see all my VMs using virsh list --all (I had to set a default_uri in .config/libvirt/libvirt.conf to get any non-empty output). Trying from another machine (say target hostname is HOSTX, username testuser) with virsh -c qemu+ssh://testusers@HOSTX/system always results in >> error: failed to connect to the hypervisor error: internal error: >> received hangup / error event on socket I see this in /var/log/messages (lines will be wrapped in the mail, sorry): > 2014-03-24T21:25:55.023525+01:00 HOSTX sshd[3509]: Accepted > publickey for testuser from 192.168.79.8 port 54176 ssh2 > 2014-03-24T21:25:55.024744+01:00 HOSTX sshd[3509]: > pam_unix(sshd:session): session opened for user testuser by (uid=0) > 2014-03-24T21:25:55.025717+01:00 HOSTX kernel: [ 3162.225673] > type=1006 audit(1395692755.024:69): pid=3509 uid=0 old > auid=4294967295 new auid=1000 old ses=4294967295 new ses=24 res=1 > 2014-03-24T21:25:55.027501+01:00 HOSTX systemd[1]: Starting Session > 24 of user testuser. 2014-03-24T21:25:55.027918+01:00 HOSTX > systemd-logind[703]: New session 24 of user testuser. > 2014-03-24T21:25:55.028213+01:00 HOSTX systemd[1]: Started Session > 24 of user testuser. 2014-03-24T21:25:55.081719+01:00 HOSTX > sshd[3511]: Received disconnect from 192.168.79.8: 11: disconnected > by user 2014-03-24T21:25:55.082095+01:00 HOSTX sshd[3509]: > pam_unix(sshd:session): session closed for user testuser > 2014-03-24T21:25:55.086268+01:00 HOSTX systemd-logind[703]: Removed > session 24. I am curious about the 'Received disconnect from 192.168.79.8: 11: disconnected by user' thing. How can I fix this? Or how to debug this? How can I get more info from libvirt? /var/log/kvm is an empty directory, /var/log/libvirt/qemu just has logs for the different VMs. I have already tried to get polkit running, although I think it should not be used due to the auth_unix_rw = "none". Any help will be appreciated. Versions: openSUSE TUmbleweed (which is openSUSE 13.1 with some more recent software) qemu-kvm is version 1.7.90-19.1.x86_64 libvirt is 1.1.2 The following packes are installed: libvirt-1.1.2-2.18.3.x86_64 libvirt-client-1.1.2-2.18.3.x86_64 libvirt-daemon-1.1.2-2.18.3.x86_64 libvirt-daemon-config-network-1.1.2-2.18.3.x86_64 libvirt-daemon-config-nwfilter-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-interface-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-libxl-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-lxc-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-network-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-nodedev-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-nwfilter-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-qemu-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-secret-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-storage-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-uml-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-vbox-1.1.2-2.18.3.x86_64 libvirt-daemon-driver-xen-1.1.2-2.18.3.x86_64 libvirt-daemon-lxc-1.1.2-2.18.3.x86_64 libvirt-daemon-qemu-1.1.2-2.18.3.x86_64 libvirt-doc-1.1.2-2.18.3.x86_64 libvirt-glib-1_0-0-0.1.7-2.1.3.x86_64 libvirt-python-1.1.2-2.18.3.x86_64 Regards, Johannes - -- Have you ever noticed that the Klingons are all speaking unix? 'Grep ls awk chmod.'' 'Mknod ksh tar imap.' 'Wall fsck yacc!' (that last is obviously a curse of some sort). Gandalf Parker -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with SeaMonkey - http://www.enigmail.net/ iEYEARECAAYFAlM0CU8ACgkQzi3gQ/xETbKKyQCfVQ4XviDo5+92ivOIgx4lXIty 980An28BKHjfwZiRh5T2Yyj3wHwOLrYJ =ujio -----END PGP SIGNATURE-----

1 2