- Users - libvirt List Archives

[libvirt-users] Undoing domiftune changes
by Brian Rak 18 Mar '14

18 Mar '14

If I apply some network limits with domiftune, how do I undo them (without restarting the VM?) So let's say I did: # virsh domiftune SRVID8736 vnet0 inbound.average: 0 inbound.peak : 0 inbound.burst : 0 outbound.average: 0 outbound.peak : 0 outbound.burst : 0 # virsh domiftune SRVID8736 vnet0 --outbound 1000,1000,1000 How I do revert this back to what it was before I set it? If I try "--outbound 0,0,0" it fails, saying "error: outbound average is mandatory"

2 1

[libvirt-users] a question on vCPU setting for lxc
by WANG Cheng D 18 Mar '14

18 Mar '14

Dear all, I am not clear about the 'vcpu' element for CPU allocation. I allocated 1 vCPU to my container, after I started the container, I ran 4 computation-intensive tasks on the container. And I found all the 4 physical core are 100% used (my host has 4 physical cores and no other application ran on the host except the container). That is, all available cores were used by the container. I want to know how to give a hard limitation for CPU usage of a container. So I don't understand what 'vcpu' setting can be used for. I know that another CPU allocation element 'shares' can also be used, but this elements only give a relative quota. If new containers are started, the CPU quota for the already started containers will change. Regards, Cheng

2 2

[libvirt-users] Sys::Virt integration into other event loops
by Scott Wiersdorf 17 Mar '14

17 Mar '14

Hi all, I’m trying to integrate Perl’s Sys::Virt into an already existing AnyEvent program. I’m accustomed to writing things like this: use EV; use AnyEvent; use AnyEvent::Handle; my $h = AnyEvent::Handle->new(fh => $fh, …); $h->on_read(sub { … }); EV::run; ## start the event loop I can add some code in the on_read() handler and every time the $fh has something to read, it will fire off. I’d like to do something similar with Sys::Virt, but I can’t seem to wrap my head around its event system. The only examples I can find are the ones included in Sys::Virt source, which consist of a series of run_once() calls, or a while loop around run_default(). Does anyone have any idea how I can make this play nicely with an existing event loop such as EV, or even to fire off ($cv->send) an AnyEvent condvar when the event I set in domain_event_register_any() triggers? Our current solution involves setting an AnyEvent->timer, which periodically fires off and then does a run_once() to see if any events happened during the period, but we’d like something more responsive (i.e., triggers when the event actually occurs) and less poll-y feeling. I’d appreciate any advice you may have—thanks! Scott

2 3

[libvirt-users] Problem in getting memory statistics
by Pasquale Dir 17 Mar '14

17 Mar '14

Hello, I need to get how much memory is used by a guest system, in order to implement some monitoring function which tells me if the system is overstressed. I am currently using java apis and the binding which was suggested to me was Domain.memoryStats(); This is a binding to int virDomainMemoryStats (virDomainPtr<http://libvirt.org/html/libvirt-libvirt.html#virDomainPtr>dom, virDomainMemoryStatPtr<http://libvirt.org/html/libvirt-libvirt.html#virDomainMemoryStatPtr>stats, unsigned int nr_stats, unsigned int flags). Problem is that it returns me just tags 0,6 and 7. Looking at the documentation I see they are not what I am looking for...I would rather need 4 (VIR_DOMAIN_MEMORY_STAT_UNUSED) and 5 (VIR_DOMAIN_MEMORY_STAT_AVAILABLE). Is there a way to set them?

2 3

Re: [libvirt-users] Problem in getting memory statistics
by Claudio Bley 17 Mar '14

17 Mar '14

[please don't top post on technical lists. it is easier to follow conversation and less likely you forget answering a question if you reply to the questions in line] [please keep the conversation on the list - I've re-added libvirt-users] At Sat, 15 Mar 2014 10:15:15 +0100, Pasquale Dir wrote: > > I am on a kubuntu 13.10 x64, qemu version 1.5.0, guest uses kvm as emulator > (1.5.0 as well), libvirt version is 1.1.1. > I don't know what is a balloon driver...so I can't tell the version of it. If you're using Linux inside your guest, just run "modinfo virtio_balloon" and post the output. > 2014-03-14 12:00 GMT+01:00 Claudio Bley <cbley(a)av-test.de>: > > > At Wed, 12 Mar 2014 13:36:50 +0100, > > Pasquale Dir wrote: > > > > > > Problem is that it returns me just tags 0,6 and 7. > > > Looking at the documentation I see they are not what I am looking for...I > > > would rather need 4 (VIR_DOMAIN_MEMORY_STAT_UNUSED) and 5 > > > (VIR_DOMAIN_MEMORY_STAT_AVAILABLE). > > > > The information available depends on the kind of hypervisor you're > > using and which guest OS you have. > > > > I guess you're using qemu. Which version are you using? > > > What guest OS do you use? Which version? Er, you missed this question...

1 0

[libvirt-users] How assign static ip using NAT?
by Peng Yu 15 Mar '14

15 Mar '14

Hi, I have the following configuration in the xml of the guest. But if I set a static ip on the guest, then the guest can not access the outside network. I don't find an example on how to set a static IP address for the guest. Could anybody know how to modify the following xml code to do so? Thanks. 41 <interface type='network'> 42 <mac address='52:54:00:6c:a7:6d'/> 43 <source network='default'/> 44 <model type='virtio'/> 45 <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> 46 </interface> -- Regards, Peng

2 1

[libvirt-users] CfP 9th Workshop on Virtualization in High-Performance Cloud Computing (VHPC '14)
by VHPC 14 15 Mar '14

15 Mar '14

we apologize if you receive multiple copies of this CfP ================================================================= CALL FOR PAPERS 9th Workshop on Virtualization in High-Performance Cloud Computing (VHPC '14) held in conjunction with Euro-Par 2014, August 25-29, Porto, Portugal ================================================================= Date: August 26, 2014 Workshop URL: http://vhpc.org Paper Submission Deadline: May 30, 2014 CALL FOR PAPERS Virtualization technologies constitute a key enabling factor for flexible resource management in modern data centers, and particularly in cloud environments. Cloud 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 NICs/HBAs 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; the increasingly adopted paradigm of Software-Defined Networking (SDN) promises to extend this flexibility to the control and data planes of network paths. 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). What is more, in data communications Network Function Virtualization (NFV) is becoming a key technology enabling a shift from supplying hardware-based network functions, to providing them in a software-based and elastic way. In conjunction with (public and private) cloud technologies, NFV may be used for constructing the foundation for cost-effective network functions that can easily and seamlessly adapt to demand, still keeping their major carrier-grade characteristics in terms of QoS and reliability. 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: - Management, deployment and monitoring of virtualized environments - Language-process virtual machines - Performance monitoring for virtualized/cloud workloads - Virtual machine monitor platforms - Topology management and optimization for distributed virtualized applications - Paravirtualized I/O - Improving I/O and network virtualization including use of RDMA, Infiniband, PCIe - Improving performance in VM access to GPUs, GPU clusters, GP-GPUs - HPC storage virtualization - Virtualized systems for big-data and analytics workloads - Optimizations and enhancements to OS virtualization support - Improving OS-level virtualization and its integration within cloud management - Performance modelling for virtualized/cloud applications - Heterogeneous virtualized environments - Network virtualization - Software defined networking - Network function virtualization - Hypervisor and network virtualization QoS and SLAs - Cloudbursting - Evolved European grid architectures including such based on network virtualization - Workload characterization for VM-based environments - Optimized communication libraries/protocols in the cloud - System and process/bytecode VM convergence - Cloud frameworks and APIs - Checkpointing/migration of VM-based large compute jobs - Job scheduling/control/policy with VMs - Instrumentation interfaces and languages - VMM performance (auto-)tuning on various load types - Cloud reliability, fault-tolerance, and security - Research, industrial and educational use cases - Virtualization in cloud, cluster and grid environments - Cross-layer VM optimizations - Cloud HPC use cases including optimizations - Services in cloud HPC - Hypervisor extensions and tools for cluster and grid computing - Cluster provisioning in the cloud - Performance and cost modelling - Languages for describing highly-distributed compute jobs - VM cloud and cluster distribution algorithms, load balancing - Instrumentation interfaces and languages - Energy-aware virtualization Important Dates Rolling Paper registration May 30, 2014 - Full paper submission July 4, 2014 - Acceptance notification October 3, 2014 - Camera-ready version due August 26, 2014 - Workshop Date TPC CHAIR Michael Alexander (chair), TU Wien, Austria Anastassios Nanos (co-chair), NTUA, Greece Tommaso Cucinotta (co-chair), Bell Labs, Dublin, Ireland PROGRAM COMMITTEE Costas Bekas, IBM Jakob Blomer, CERN Roberto Canonico, University of Napoli Federico II, Italy Paolo Costa, MS Research Cambridge, England Jorge Ejarque Artigas, Barcelona Supercomputing Center, Spain William Gardner, University of Guelph, USA Balazs Gerofi, University of Tokyo, Japan Krishna Kant, Temple University, USA Romeo Kinzler, IBM Nectarios Koziris, National Technical University of Athens, Greece Giuseppe Lettieri, University of Pisa, Italy Jean-Marc Menaud, Ecole des Mines de Nantes, France Christine Morin, INRIA, France Dimitrios Nikolopoulos, Queen's University of Belfast, UK Herbert Poetzl, VServer, Austria Luigi Rizzo, University of Pisa, Italy Josh Simons, VMWare, USA Borja Sotomayor, University of Chicago, USA Vangelis Tasoulas, Simula Research Lab, Norway Yoshio Turner, HP Labs, USA Kurt Tutschku, Blekinge Institute of Technology, Sweden 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 EasyChair Abstract Submission Link: https://www.easychair.org/conferences/?conf=europar2014ws GENERAL INFORMATION The workshop is one day in length and will be held in conjunction with Euro-Par 2014, 25-29 August, Porto, Portugal

1 0

[libvirt-users] shutdown all domains
by Peng Yu 15 Mar '14

15 Mar '14

Hi, The following shows a script to shutdown all domains. Is this still not available in libvirt-bin? If not, will it added in the future? Thanks. http://lists.centos.org/pipermail/centos-virt/2009-November/001358.html -- Regards, Peng

2 1

Re: [libvirt-users] Scheduler Parameters
by Faruk Caglar 14 Mar '14

14 Mar '14

Hi All, I have emailed the message below to Arnaud, but it got failed. I will need to have get set scheduling parameters (including timeslice, ratelimit, weight, cap for Xen) thru C# libvirt bindings. is there any plan to have those wrappers soon? Thanks Faruk On Fri, Mar 14, 2014 at 4:06 PM, Faruk Caglar <faruk.caglar(a)vanderbilt.edu>wrote: > > Hi Arnaud, > > Thanks for all your work for the libvirt C# bindings. > > Recently I have started to use libvirt C# bindings for my doctoral > research. However, > I have noticed that it currently does not support the get/set scheduler > parameters. > > Do you have any plan to implement those in a near future? > > Thanks > Faruk > Vanderbilt University > >

1 0

[libvirt-users] questions on clock catchup
by Jincheng Miao 13 Mar '14

13 Mar '14

In http://libvirt.org/formatdomain.html#elementsTime , we could find there are three attributes of catchup tickpolicy: limit, threshold and slew. " The catchup element has three optional attributes, each a positive integer. The attributes are threshold, slew, and limit. " The xml format likes: <clock offset='utc'> <timer name='rtc' tickpolicy='catchup'> <catchup slew='123'/> or <catchup threshold='123'/> or <catchup limit='123'/> </timer> </clock> But there is no further explanation. Does anyone know what's the meaning of these attributes ? Best wishes, Jincheng Miao

5 4