March 2015 - Users - libvirt List Archives

[libvirt-users] pty pairing host - guest
by Neel Basu 19 Mar '15

19 Mar '15

HI!, I am using virt-manager GUI. There I can add a `Console Device` through Add Hardware. I can add a virtio pty WITHOUT restarting the VM. also I get a pty allocated to host machine say `/dev/pty/M` My expectation is this `host:/dev/pty/M` is connected with a `guest:/dev/ttyN`. so If I open minicom or screen and write to `host:/dev/pty/M` it will be propagated to 'guest:/dev/ttyN'. Is my expectation correct ? If it is correct how will I know what is the guest tty allocated against this host pty ? This is what I've tried: dumpxml returns <console type='pty' tty='/dev/pts/1'> <source path='/dev/pts/1'/> <target type='serial' port='0'/> <alias name='serial0'/> </console> I expected host:/dev/pts/1 is paired with guest:/dev/ttyS0. so I opened screen/cat/echo on both sides but could not exchange messages ------ Similar to console device I can add a Serial device but that requires a reboot of the guest VM. However I tried that though I prefer to do it without a guest restart. <serial type='pty'> <source path='/dev/pts/1'/> <target port='0'/> <alias name='serial0'/> </serial> here also I expected host:/dev/pts/1 is paired with /dev/ttyS0 but couldn't communicate. My sole requirement is to have a tty pair in host and guest that can be used to communicate with each other. I am using qemu-kvm. Thank You.

2 1

[libvirt-users] how to get disk snapshot size
by Yitao Jiang 16 Mar '15

16 Mar '15

Hi,guys I wanna to get the disk snapshot size, but found nothing libvirt commands related,except qemu-img. I create two disk snapshots, but nothing return its' size [root@cskvm01 qcow2]# qemu-img info /mnt/e6758700-af68-3c06-ade3-53f5f9b93507/e2cf6551-0d2c-4382-a86c-8ba633954ff2 image: /mnt/e6758700-af68-3c06-ade3-53f5f9b93507/e2cf6551-0d2c-4382-a86c-8ba633954ff2 file format: qcow2 virtual size: 5.0G (5368709120 bytes) disk size: 1.0G cluster_size: 65536 Snapshot list: ID TAG VM SIZE DATE VM CLOCK 1 4fc42d73-8257-42bd-8807-81700fd3c689 0 2015-03-15 21:05:55 01:05:53.583 2 2fd6aeab-cb26-446d-b1c6-e8d70d33f651 0 2015-03-15 21:50:35 00:00:00.000 After wrote data to the disk, then create another snapshot [root@cskvm01 qcow2]# qemu-img info /mnt/e6758700-af68-3c06-ade3-53f5f9b93507/e2cf6551-0d2c-4382-a86c-8ba633954ff2 image: /mnt/e6758700-af68-3c06-ade3-53f5f9b93507/e2cf6551-0d2c-4382-a86c-8ba633954ff2 file format: qcow2 virtual size: 5.0G (5368709120 bytes) disk size: 3.9G cluster_size: 65536 Snapshot list: ID TAG VM SIZE DATE VM CLOCK 1 4fc42d73-8257-42bd-8807-81700fd3c689 0 2015-03-15 21:05:55 01:05:53.583 2 2fd6aeab-cb26-446d-b1c6-e8d70d33f651 0 2015-03-15 21:50:35 00:00:00.000 3 63815565-3a06-4366-a1b3-bfeb9c4a07b4 0 2015-03-15 22:22:43 00:00:00.000 VM SIZE column still show 0 Here is my environment qemu-img version 0.12.1, Copyright (c) 2004-2008 Fabrice Bellard libvirtd (libvirt) 0.10.2 CentOS release 6.5 (Final) 2.6.32-431.el6.x86_64 --- Thanks, Yitao(依涛姜) jiangyt.github.io

3 3

[libvirt-users] Processor usage of qemu process.
by Dominique Ramaekers 15 Mar '15

15 Mar '15

I have been using libvirt for a while now with some linux guest installed. And everything has been working great. I've got a nice new (used) HP virtual host with 12 x dual core and 48Gb mem. My Windows servers are getting old, so I found it was time to take the next step and also virtualise my Windows systems. Now I've got two Windows guests on my new host: - A Windows 8.1 which runs a Autodesk Job Processor - A Windows Server 2012 R2 which runs a Pervasive SQL, Autodesk Vault Server (MSSQL, IIS), Autodesk license server,... The first is constanly using 30% host-CPU and the second one is using 50% host-CPU. Both guests are using <5% guest CPU... (Host system load is between 0.1 and 0.3...) Is it normal Windows guests take up 30% to 50% host CPU resources? The Windows 8.1 is actualy a simpel PC with almost no activity... Version: QEMU emulator version 2.0.0 (Debian 2.0.0+dfsg-2ubuntu1.10), Copyright (c) 2003-2008 Fabrice Bellard

3 13

[libvirt-users] How to run qemu with root permission from libvirt.
by Edward Young 15 Mar '15

15 Mar '15

Hi all, My platform is qemu 1.5.1 and libvirt 1.2.0. How to lauch the qemu with root permission, since in the qemu I would like to access the hardware directly. I actually check the runing /usr/bin/kvm process, it indicated that this process is from a root user, but it still failed to act as with root permission. Thanks! Ed

2 1

[libvirt-users] lxc-enter-namespace support in Python API
by Florian Haas 13 Mar '15

13 Mar '15

Hello everyone, referring back to https://www.redhat.com/archives/libvirt-users/2013-August/msg00107.html, in which a user discovered that there is no equivalent to "virsh -c lxc:/// lxc-enter-namespace" in the Python API. Has that ever changed? In that thread Daniel suggested that the user file a bug, but that apparently never happened. Why am I interested in this? Ansible currently supports a somewhat limited libvirt_lxc connection driver which uses lxc-enter-namespace to access LXC containers without SSH (https://github.com/ansible/ansible/blob/devel/lib/ansible/runner/connection…) That connection driver, written in Python like the rest of Ansible, currently has to use jump through subprocess hoops with virsh to enter the namespace, which is slow and doesn't support pipelining. Being able to use the native Python API would help greatly. In addition, if it were to use the Python API instead, the connection plugin could also be extended to use lxc+ssh:// connection types so Ansible users could potentially manage container hosts _and_ all their containers from a single playbook, just so long as they have SSH access to the host. Does anyone know whether that extension of the API is being / has been addressed? Cheers, Florian

1 0

[libvirt-users] lxc-enter-namespace support in Python API
by Florian Haas 13 Mar '15

13 Mar '15

Hello everyone, referring back to https://www.redhat.com/archives/libvirt-users/2013-August/msg00107.html, in which a user discovered that there is no equivalent to "virsh -c lxc:/// lxc-enter-namespace" in the Python API. Has that ever changed? In that thread Daniel suggested that the user file a bug, but that apparently never happened. Why am I interested in this? Ansible currently supports a somewhat limited libvirt_lxc connection driver which uses lxc-enter-namespace to access LXC containers without SSH (https://github.com/ansible/ansible/blob/devel/lib/ansible/runner/connection…) That connection driver, written in Python like the rest of Ansible, currently has to use jump through subprocess hoops with virsh to enter the namespace, which is slow and doesn't support pipelining. Being able to use the native Python API would help greatly. In addition, if it were to use the Python API instead, the connection plugin could also be extended to use lxc+ssh:// connection types so Ansible users could potentially manage container hosts _and_ all their containers from a single playbook, just so long as they have SSH access to the host. Does anyone know whether that extension of the API is being / has been addressed? Cheers, Florian

1 0

[libvirt-users] vm live storage migration failure.
by Edward Young 11 Mar '15

11 Mar '15

Hi all, When I migrate a live vm with two virtual disk images from one node to another. Bothe of the nodes are in the same LAN and there is no shared storage. I run the following command and get the output blew: yyang@node1:~$ sudo virsh migrate --live --persistent --copy-storage-all --unsafe --verbose vm1 qemu+ssh://192.168.1.3/system root(a)192.168.1.3's password: root(a)192.168.1.3's password: root(a)192.168.1.3's password: error: Cannot recv data: Permission denied, please try again. Permission denied, please try again. Permission denied (publickey,password).: Connection reset by peer I'm sure the root passwd I entered is correct. And When I run the following command, it works. I can use qemu+ssh to connect to a remote libvirtd daemon. yyang@node2:~$ sudo virsh Welcome to virsh, the virtualization interactive terminal. Type: 'help' for help with commands 'quit' to quit virsh # connect qemu+ssh://192.168.1.2/system root(a)192.168.1.2's password: virsh # list Id Name State ---------------------------------------------------- 6 vm1 running 7 vm2 running virsh # Any suggestions about this issue? Thanks a lot! Ed

2 1

[libvirt-users] CfP 10th Workshop on Virtualization in High-Performance Cloud Computing (VHPC '15)
by VHPC 15 11 Mar '15

11 Mar '15

================================================================= CALL FOR PAPERS 10th Workshop on Virtualization in High-Performance Cloud Computing (VHPC '15) held in conjunction with Euro-Par 2015, August 24-28, Vienna, Austria (Springer LNCS) ================================================================= Date: August 25, 2015 Workshop URL: http://vhpc.org Paper Submission Deadline: May 22, 2015 CALL FOR PAPERS Virtualization technologies constitute a key enabling factor for flexible resource management in modern data centers, cloud environments, and increasingly in HPC as well. Providers need to dynamically manage complex infrastructures in a seamless fashion for varying workloads and hosted applications, independently of the customers deploying software or users submitting highly dynamic and heterogeneous workloads. Thanks to virtualization, we have the ability to manage vast computing and networking resources dynamically and close to the marginal cost of providing the services, which is unprecedented in the history of scientific and commercial computing. Various virtualization technologies contribute to the overall picture in different ways: machine virtualization, with its capability to enable consolidation of multiple under-utilized servers with heterogeneous software and operating systems (OSes), and its capability to live-migrate a fully operating virtual machine (VM) with a very short downtime, enables novel and dynamic ways to manage physical servers; OS-level virtualization, with its capability to isolate multiple user-space environments and to allow for their co-existence within the same OS kernel, promises to provide many of the advantages of machine virtualization with high levels of responsiveness and performance; I/O Virtualization allows physical network adapters to take traffic from multiple VMs; network virtualization, with its capability to create logical network overlays that are independent of the underlying physical topology and IP addressing, provides the fundamental ground on top of which evolved network services can be realized with an unprecedented level of dynamicity and flexibility; These technologies have to be inter-mixed and integrated in an intelligent way, to support workloads that are increasingly demanding in terms of absolute performance, responsiveness and interactivity, and have to respect well-specified Service- Level Agreements (SLAs), as needed for industrial-grade provided services. Indeed, among emerging and increasingly interesting application domains for virtualization, we can find big-data application workloads in cloud infrastructures, interactive and real-time multimedia services in the cloud, including real-time big-data streaming platforms such as used in real-time analytics supporting nowadays a plethora of application domains. Distributed cloud infrastructures promise to offer unprecedented responsiveness levels for hosted applications, but that is only possible if the underlying virtualization technologies can overcome most of the latency impairments typical of current virtualized infrastructures (e.g., far worse tail-latency). The Workshop on Virtualization in High-Performance Cloud Computing (VHPC) aims to bring together researchers and industrial practitioners facing the challenges posed by virtualization in order to foster discussion, collaboration, mutual exchange of knowledge and experience, enabling research to ultimately provide novel solutions for virtualized computing systems of tomorrow. The workshop will be one day in length, composed of 20 min paper presentations, each followed by 10 min discussion sections, and lightning talks, limited to 5 minutes. Presentations may be accompanied by interactive demonstrations. TOPICS Topics of interest include, but are not limited to: - Virtualization in supercomputing environments, HPC clusters, cloud HPC and grids - Optimizations of virtual machine monitor platforms, hypervisors and OS-level virtualization - Hypervisor and network virtualization QoS and SLAs - Cloud based network and system management for SDN and NFV - Management, deployment and monitoring of virtualized environments - Performance measurement, modelling and monitoring of virtualized/cloud workloads - Programming models for virtualized environments - Cloud reliability, fault-tolerance, high-availability and security - Heterogeneous virtualized environments, virtualized accelerators, GPUs and co-processors - Optimized communication libraries/protocols in the cloud and for HPC in the cloud - Topology management and optimization for distributed virtualized applications - Cluster provisioning in the cloud and cloud bursting - Adaptation of emerging HPC technologies (high performance networks, RDMA, etc..) - I/O and storage virtualization, virtualization aware file systems - Job scheduling/control/policy in virtualized environments - Checkpointing and migration of VM-based large compute jobs - Cloud frameworks and APIs - Energy-efficient / power-aware virtualization Important Dates April 29, 2015 - Abstract registration May 22, 2015 - Full paper submission June 19, 2014 - Acceptance notification October 2, 2014 - Camera-ready version due August 25, 2014 - Workshop Date TPC CHAIR Michael Alexander (chair), TU Wien, Austria Anastassios Nanos (co-chair), NTUA, Greece Balazs Gerofi (co-chair), RIKEN Advanced Institute for Computational Science, Japan PROGRAM COMMITTEE Stergios Anastasiadis, University of Ioannina, Greece Costas Bekas, IBM Zurich Research Laboratory, Switzerland Jakob Blomer, CERN Ron Brightwell, Sandia National Laboratories, USA Roberto Canonico, University of Napoli Federico II, Italy Julian Chesterfield, OnApp, UK Patrick Dreher, MIT, USA William Gardner, University of Guelph, Canada Kyle Hale, Northwestern University, USA Marcus Hardt, Karlsruhe Institute of Technology, Germany Iftekhar Hussain, Infinera, USA Krishna Kant, Temple University, USA Eiji Kawai, National Institute of Information and Communications Technology, Japan Romeo Kinzler, IBM, Switzerland Kornilios Kourtis, ETH, Switzerland Nectarios Koziris, National Technical University of Athens, Greece Massimo Lamanna, CERN Che-Rung Roger Lee, National Tsing Hua University, Taiwan Helge Meinhard, CERN Jean-Marc Menaud, Ecole des Mines de Nantes France Christine Morin, INRIA, France Amer Qouneh, University of Florida, USA Seetharami Seelam, IBM Watson Research Center, USA Josh Simons, VMWare, USA Borja Sotomayor, University of Chicago, USA Kurt Tutschku, Blekinge Institute of Technology, Sweden Yasuhiro Watashiba, Osaka University, Japan Chao-Tung Yang, Tunghai University, Taiwan PAPER SUBMISSION-PUBLICATION Papers submitted to the workshop will be reviewed by at least two members of the program committee and external reviewers. Submissions should include abstract, key words, the e-mail address of the corresponding author, and must not exceed 10 pages, including tables and figures at a main font size no smaller than 11 point. Submission of a paper should be regarded as a commitment that, should the paper be accepted, at least one of the authors will register and attend the conference to present the work. Accepted papers will be published in the Springer LNCS series - the format must be according to the Springer LNCS Style. Initial submissions are in PDF; authors of accepted papers will be requested to provide source files. Format Guidelines: http://www.springer.de/comp/lncs/authors.html Submission Link: https://easychair.org/conferences/?conf=europar2015ws GENERAL INFORMATION The workshop is one day in length and will be held in conjunction with Euro-Par 2015, 24-28 August, Vienna, Austria

1 0

[libvirt-users] Issues with XML validation after upgrade to 1.2.12
by Brian Rak 10 Mar '15

10 Mar '15

After we upgraded to 1.2.12, we've been having issues with libvirt... it complains that our formerly valid guest definitions are now invalid: error: Failed to start domain XXXX error: internal error: Cannot instantiate filter due to unresolvable variables or unavailable list elements: DHCPSERVER We looked into this, and found that it's the XML validation that's failing: # xmllint --noout --relaxng "/share/libvirt/schemas/domain.rng" XXXX.xml --recover Relax-NG validity error : Extra element devices in interleave test.xml:1: element domain: Relax-NG validity error : Element domain failed to validate content test.xml fails to validate And, a minimal domain XML to reproduce it (this won't boot, but it shows the issue): <domain type='kvm' id='65'> <name>XXXX</name> <uuid>b602b5f2-b9d7-43bd-a949-acc7eeeb9f8f</uuid> <memory unit='KiB'>1048576</memory> <devices> <interface type='bridge'> <filterref filter='myfilter'> <parameter name='CTRL_IP_LEARNING' value='none'/> <parameter name='DHCPSERVER' value='104.156.226.10'/> <parameter name='IP' value='104.207.129.11'/> <parameter name='IP6_ADDR' value='2001:19f0:300:2102::'/> <parameter name='IP6_MASK' value='64'/> </filterref> </interface> </devices> </domain> The cause seems to be having multiple parameters in a <filterref> block. We applied the following patch to fix it: diff -ur src_clean/docs/schemas/domaincommon.rng src/docs/schemas/domaincommon.rng --- src_clean/docs/schemas/domaincommon.rng 2015-01-23 06:46:24.000000000 -0500 +++ src/docs/schemas/domaincommon.rng 2015-03-10 11:30:42.057441342 -0400 @@ -4468,6 +4468,7 @@ <data type="NCName"/> </attribute> <optional> + <zeroOrMore> <element name="parameter"> <attribute name="name"> <ref name="filter-param-name"/> @@ -4476,6 +4477,7 @@ <ref name="filter-param-value"/> </attribute> </element> + </zeroOrMore> </optional> </define>

2 1

[libvirt-users] Unable to start sandbox: Kernel module dir /lib/modules/3.18.5-x86_64-linode52/kernel does not exist
by Adam Smith 09 Mar '15

09 Mar '15

Dear all, I have been trying to set up the set up Libvirt Sandbox without success. I want to use virt-sandbox in order to run untrusted programs in a secure environment. I am had no knowledge about virtualization until a couple of days ago, so I am probably doing something wrong. The scenario is the following: Linode instance. OS that I have tried: Ubuntu 14.04, Ubuntu 14, Fedora 21. Both compiling from source and installing the pre-compiled packages. But I always reach the same error: """ $ virt-sandbox -c qemu:///session /bin/date Unable to start sandbox: Kernel module dir /lib/modules/3.18.5-x86_64-linode52/kernel does not exist """ I have been told by the guys of Linode that: "The kernels we use are completely compiled and do not utilize modules. In addition, the kernels are loaded from the host rather than the /boot directory" Any hints to solve this issue? Is the only solution to compile my own kernel? Also, if I decide to use a service like Linode, AWS, Digital Ocean...then the server that I would be using would be already a virtual server. Is it a problem to run virt-sandbox within a server which is already a virtual server? Thanks a lot!

2 1