June 2015 - Users - libvirt List Archives

Re: [libvirt-users] [ovirt-users] Bug in Snapshot Removing
by Soeren Malchow 11 Jun '15

11 Jun '15

Small addition again: This error shows up in the log while removing snapshots WITHOUT rendering the Vms unresponsive — Jun 01 01:33:45 mc-dc3ham-compute-02-live.mc.mcon.net libvirtd[1657]: Timed out during operation: cannot acquire state change lock Jun 01 01:33:45 mc-dc3ham-compute-02-live.mc.mcon.net vdsm[6839]: vdsm vm.Vm ERROR vmId=`56848f4a-cd73-4eda-bf79-7eb80ae569a9`::Error getting block job info Traceback (most recent call last): File "/usr/share/vdsm/virt/vm.py", line 5759, in queryBlockJobs… — From: Soeren Malchow <soeren.malchow(a)mcon.net<mailto:soeren.malchow@mcon.net>> Date: Monday 1 June 2015 00:56 To: "libvirt-users(a)redhat.com<mailto:libvirt-users@redhat.com>" <libvirt-users(a)redhat.com<mailto:libvirt-users@redhat.com>>, users <users(a)ovirt.org<mailto:users@ovirt.org>> Subject: [ovirt-users] Bug in Snapshot Removing Dear all I am not sure if the mail just did not get any attention between all the mails and this time it is also going to the libvirt mailing list. I am experiencing a problem with VM becoming unresponsive when removing Snapshots (Live Merge) and i think there is a serious problem. Here are the previous mails, http://lists.ovirt.org/pipermail/users/2015-May/033083.html The problem is on a system with everything on the latest version, CentOS 7.1 and ovirt 3.5.2.1 all upgrades applied. This Problem did NOT exist before upgrading to CentOS 7.1 with an environment running ovirt 3.5.0 and 3.5.1 and Fedora 20 with the libvirt-preview repo activated. I think this is a bug in libvirt, not ovirt itself, but i am not sure. The actual file throwing the exception is in VDSM (/usr/share/vdsm/virt/vm.py, line 697). We are very willing to help, test and supply log files in anyway we can. Regards Soeren

3 12

[libvirt-users] SSH into LXC and install files via JAVA API
by Anu Mercian 10 Jun '15

10 Jun '15

Hello everyone, I am trying to use libvirt's JAVA API to ssh into a domain and install some files, for example, dhcpd. Using virsh, its pretty straight forward using the command line: virsh -c lxc:/// create myguest.xml virsh -c lxc:/// console myguest virsh -c lxc:/// lxc-enter-namespace myguest -- sudo yum install dhcpd But in order to incorporate the above functionality in a JAVA API, I am not really sure what to do. The following command helps me ssh into a system, but I am not sure how to send remote commands. conn = Connect("lxc+ssh:///") Please can someone help me out with this? Thank you, Anu

1 0

[libvirt-users] how do I use the same secret with the same ID on all libvrit/qemu/kvm hosts?
by Jelle de Jong 09 Jun '15

09 Jun '15

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello everybody, Every time I use [virsh secret-define --file secret.xml] I get an unique id back and I use this id in my vm guests xml configs, but this is preventing me from live migrating between kvm host. Maybe im using the wrong methods, but i normaly dumpxml the config and define it in the other kvm hosts (scripted) but now this other host has an other unique id and the kvm guests fails so I'm guessing I need to use the same secret id betwen all my kvm hosts but how do i do this? Kind regards, Jelle de Jong -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iJwEAQECAAYFAlV3Pa0ACgkQ1WclBW9j5HmSsAP/ZqG0kqf+pQPD1Jng+qIA1Ih8 qEabjpZyBAz5AITP6RpeopP+ZAyXIlex4xmSGPBNZPT2OEC9J4jA0AcC65MiG44f 1ma2C7VckoPoONO0D9q8JLwqG8fFVqAFZBTszBz5sj7qzEPiIC5/g6b0czf8QUno b9LwvViwGCPgfETBkXU= =zT/0 -----END PGP SIGNATURE-----

1 1

[libvirt-users] LXCs using Libvirt
by Anu Mercian 09 Jun '15

09 Jun '15

Hi all, I am trying to bring up lxc's using Libvirt JAVA API. I do the following: conn = Connect("lxc:\\\") xml_config = "<domain> .... <\domain>" dom = conn.domainCreateXML(xml_config, 0) dom.create() Using the above code, I am able to start a domain but the domain is not visible within the "lxc" list, i.e when I use the command "lxc-ls --fancy", the domain that was created cannot be viewed, although, when I use the function conn.listDomains(), I can view the domain. As a work-around if I use the "lxc-create -t <OS> -n <name>" command followed by "conn.domainDefineXML(xml_config)", I can view the domain within "lxc-ls --fancy" Any one has observed this problem earlier and is there a way I can create the domain without using "lxc-create ..." command and still list the domain on the "lxc-ls --fancy" list? Thank you, Anu

2 2

[libvirt-users] Freeze Windows Guests For Consistent Storage Snapshots
by Payes Anand 09 Jun '15

09 Jun '15

Hi, Is it possible to freeze windows guests for a consistent storage level snapshot. I am using openstack icehouse on centos 6.6 Hypervisor: KVM Libvirt: 0.10.2 Qemu: 0.10.2 Guest OS: Windows 7 and Windows Server 2008 I was able to freeze Centos guests by issuing the command: virsh qemu-agent-command <guest_ID> '{"execute":"guest-fsfreeze-freeze"}' For CentOS guests, I enabled access between compute nodes and guests through a socket by setting metadata parameter, hw_qemu_guest_agent=yes for the guest image. And then installing qemu-guest-agent inside the guest. What steps do i have to follow for windows ? Regars, Payes

2 2

[libvirt-users] Help with live Migration
by suthershan sundralingam 08 Jun '15

08 Jun '15

Hi All, I am trying to measure live migration data. I am interested in the following data: TIME elapsed: Data processed: Data remaining: Data total: Constant pages : Normal pages : Normal data : Expected down time : What is the command i am supposed to run? Suthershan

1 0

[libvirt-users] Qemu-Libgfapi: periodical shutdown of virtual machines.
by Igor Yakovlev 08 Jun '15

08 Jun '15

Hello, Would you so kind to help me with my problem concerning libgfapi. My host operating system is Ubuntu 14.04 LTS, version of glusterfs is 3.6.2, and version of qemu is 2.0.0. In our environment guest virtual machines periodically go to power-off state 'Powered Off'(Shutdown) with ERROR in /var/log/syslog like: kernel: [5346607.988173] qemu-system-x86[29564]: segfault at 128 ip 00007f930a90b48c sp 00007f931296fd70 error 4 in qemu-system-x86_64[7f930a57f000+4b1000] Do you have any ideas?... Best regards, Igor.

2 2

[libvirt-users] (no subject)
by Akendo 08 Jun '15

08 Jun '15

2 1

[libvirt-users] agent discovery
by Stuart Kendrick 03 Jun '15

03 Jun '15

How might I discover, via SNMP, which boxes are running libvirt-snmp? Performing a walk from the top seems to skip the Libvirt MIB ... (why is this?) ~/walk$ snmpbulkwalk -c public kvmtestbox .iso | more SNMPv2-MIB::sysDescr.0 = STRING: Linux kvmtestbox.company.com 2.6.32 -358.23.2.el6.x86_64 #1 SMP Wed Oct 16 18:37:12 UTC 2013 x86_64 SNMPv2-MIB::sysObjectID.0 = OID: NET-SNMP-TC::linux EXPRESSION-MIB::sysUpTimeInstance = Timeticks: (1308205914) 151 days, 9:54:19.14 [...] SNMP-VIEW-BASED-ACM-MIB::vacmViewTreeFamilyStatus."systemview".9.1.3.6.1.2.1.25. 1.1 = No more variables left in this MIB View (It is past the end of the MIB tre e) I don't see any particular variable I could query ... no equivalent of: LIBVIRT-MIB::libvirtSupport = 1 (yes) Although I suppose a walk starting at the agent's enterprise OID works: ~/walk$ snmpbulkwalk -c public kvmtestbox enterprises.12345 | more LIBVIRT-MIB::libvirtGuestName.'.1..t.X.s>..To..' = STRING: "test1" LIBVIRT-MIB::libvirtGuestName.'7". )...L....(..' = STRING: "test2" LIBVIRT-MIB::libvirtGuestName.'D'q2...*R5y..2&.' = STRING: "test3" [...] But ... if the admin of the hosting box has changed the enterprise OID from '12345' to something more unique, then this doesn't work. How do you discover boxes which support libvirt-snmp? --sk

1 0

[libvirt-users] Missing "block-stream" command in qemu-kvm
by Dallas Graves 03 Jun '15

03 Jun '15

Hi all, I’m experiencing an interesting issue where I’m attempting to consolidate my snapshots’ backing images on my qemu-kvm install by way of the virsh blockpull and/or virsh blockcommit commands and receiving the error that my system can not locate the “block-stream” command. I can’t seem to find much related to this issue by way of my Google research outside of some Bugzilla posts, which leads me to believe that perhaps I’ve missed something during my install, maybe there is an ancillary package I need to install? I’m running qemu-kvm-0.12.1.2-2.448.el6_6.3.x86_64 with libvirt-0.10.2-29.el6_5.12.x86_64 on CentOS 6.6. If anyone has experienced this issue before, advise would be greatly appreciated. Thanks, Dallas G.

2 2

[libvirt-users] Canging bridge names on live migration
by seitan 03 Jun '15

03 Jun '15

Hello, i wonder, if there's a possibility to change a name of a shared interface in virtual machine config, while doing migration. The problem is: hypervisor1 (source) uses shared interface name "br0". hypervisor2 (target) uses shared interface name "br500". Live migration fails, because target hypervisor does not have "br0" interface. Thank you

3 2

[libvirt-users] preallocation=full Vs preallocation=metadata
by Jatin Davey 03 Jun '15

03 Jun '15

Hi All I was reading through this: http://kashyapc.com/2011/12/02/little-more-disk-io-perf-improvement-with-fa… I was basically searching for pointers on improving disk I/O. I wanted to know the purpose of preallocation=full & preallocation=metadata , What is the difference between them ? and which one would yield a better disk I/O speed ? Appreciate your responses in this regard. Thanks Jatin

2 2

[libvirt-users] Bug in Snapshot Removing
by Soeren Malchow 01 Jun '15

01 Jun '15

Dear all I am not sure if the mail just did not get any attention between all the mails and this time it is also going to the libvirt mailing list. I am experiencing a problem with VM becoming unresponsive when removing Snapshots (Live Merge) and i think there is a serious problem. Here are the previous mails, http://lists.ovirt.org/pipermail/users/2015-May/033083.html The problem is on a system with everything on the latest version, CentOS 7.1 and ovirt 3.5.2.1 all upgrades applied. This Problem did NOT exist before upgrading to CentOS 7.1 with an environment running ovirt 3.5.0 and 3.5.1 and Fedora 20 with the libvirt-preview repo activated. I think this is a bug in libvirt, not ovirt itself, but i am not sure. The actual file throwing the exception is in VDSM (/usr/share/vdsm/virt/vm.py, line 697). We are very willing to help, test and supply log files in anyway we can. Regards Soeren

2 1

[libvirt-users] snapshots and vmdk
by Boylan, Ross 01 Jun '15

01 Jun '15

Does libvirt support snapshotting when the virtual disk comes from a vmdk file? http://wiki.libvirt.org/page/Snapshots#Desired_functionality seems to say no, since it says "'virsh snapshot', which requires all disk images to be qcow2". OTOH, man virsh, http://libvirt.org/formatsnapshot.html, and http://libvirt.org/formatdomain.html#elementsDisks seem to indicate more flexibility, though I see nothing about vmdk. If there is support for vmdk, is it just for external snapshots, or do internal snapshots work? Currently running libvirt 0.9.12, qemu-kvm 1.1.2 (though I could use vmware workstation), linux kernel 3.2.0. Thanks. Ross Boylan

2 2

[libvirt-users] Block Commit: [100 %]error: failed to pivot job for disk vda (Bug filed)
by Matthew Schumacher 01 Jun '15

01 Jun '15

Hello List, Seems like blockcopy --wait doesn't work work qemu-2.2.x. I found someone else having the same problem, but it never went anywhere: https://www.redhat.com/archives/libvirt-users/2015-January/msg00045.html I went ahead and did some more debugging and created a bug report: https://bugzilla.redhat.com/show_bug.cgi?id=1210903 I'm going to do some more debugging and hopefully figure this out, but wanted it to post it here so that it gets more attention. Thanks, schu

2 4