- Users - libvirt List Archives

Re: [libvirt-users] LXC libvirt support
by Nirmal Guhan 18 Feb '11

18 Feb '11

2011/2/11 Nirmal Guhan <vavatutu(a)gmail.com>: > On Fri, Feb 11, 2011 at 9:05 PM, Osier Yang <jyang(a)redhat.com> wrote: >> 于 2011年02月12日 11:52, Nirmal Guhan 写道: >>> >>> Hi, >>> >>> Has anyone used libvirt to start/stop/manage LXC (linux containers) ? >>> Is it supported at all? Last I tried, had few issues and the reply was >>> : >> >> Yes, it's supported, though still much work to do, but the basic is >> there. >> >>> >>> /// >>> Note that libvirt lxc driver is a different implementation from lxc tools >>> and doesn't use lxc tools internally. >> >> Indeed, it doesn't invoke lxc tools. >> >> You may not be able to do what >>> >>> you are doing with lxc tools. Especially, the driver doesn't take aware >>> of files and settings inside a container at all. You have to prepare them >>> by yourself in advance >> >> libvirt.org has the example, e.g. the busybox example, I guess it could >> explain "the driver doesn't take care of files and setting inside a >> container at all, you have to prepare them by yourself in advance." >> >> Regards >> Osier > > Thanks. Will give a try. > ~nirmal > Seeing some strange behavior running fedora 12 for both host and lxc container. I do a define, start and console : [root@fedora12new lxc-libvirt-scripts]# virsh -c lxc:// console testf12new Connected to domain testf12new Escape character is ^] --> I press enter here and virsh exits. I don't get login prompt always. Sometimes I do and when I try to login it rejects my password. Apparently, the same rootfs works fine with lxc tools and I can login. Snippet from etc/event.d/tty1 : respawn exec /sbin/mingetty pts/0 I use "exec /sbin/mingetty tty1" for lxc tools and that works fine. I had to try pts/0 for libvirt since tty1 does not even get me the occasional login prompts I get with pts. Also even after virsh shutdown, I see that the container processes do not exit. Is there something else I should do? ~nirmal

3 2

[libvirt-users] altering virtual network driver iptables behavior
by Ryan Sumner 18 Feb '11

18 Feb '11

I have the need to modify the behavior of the virtual network driver's behavior and how it deals with routed networks. I'm running libvirt-0.8.3-2.fc14. According to http://libvirt.org/firewall.html, the following is automatically added to the FORWARD chain of iptables when a network type of "routed" is started up: "Allow inbound, but only to our expected subnet. Allow outbound, but only from our expected subnet. Allow traffic between guests. Deny all other inbound. Deny all other outbound. " The part of this that I need to adjust is the fact that only IPs on my subnet will be allowed in and out. I have IP addresses assigned to my guests that have static routes configured on the hypervisor to route to the local bridged interface. I have to do this due to the way the surrounding routers and switches on the network are configured to handle public IP addresses and MAC address filtering. Here is an example of my config where the public IP address on the guest machine is 1.1.2.2: My network to do the routing mode: <network> <name>local</name> <forward dev='eth0' mode='route'/> <bridge name='virbr_local' stp='on' delay='0' /> <ip address='192.168.122.1' netmask='255.255.255.0'/> </network> The network portion of the domain: <interface type='network'> <mac address='xx:xx:xx:cc:xx:xx'/> <source network='local'/> <target dev='vnet0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </interface> And finally the static route configured on the hypervisor to get the routed traffic from eth0 on the hypervisor down to the virbr_local device: # ip route add 1.1.2.2 dev virbr_local This will setup the iptables filters just as the documentation defines, but the problem is that all traffic from the guest will get REJECTED by iptables due to the source and destination IP not falling within 192.168.122.0/24. I've tried adding a custom filter into the network filter driver, but haven't had much luck. Here are some of the things that I've tried. The custom network filter. Notice that I'm using tcp, udp, and icmp specifically. I'm doing this so it will force inclusion into the iptables filtering rules rather than ebtables. <filter name='my-static-ip' chain='root'> <rule action='accept' direction='out' priority='500'> <tcp srcipaddr='$MYIP'/> </rule> <rule action='accept' direction='out' priority='500'> <udp srcipaddr='$MYIP'/> </rule> <rule action='accept' direction='out' priority='500'> <icmp srcipaddr='$MYIP'/> </rule> <rule action='accept' direction='in' priority='500'> <tcp dstipaddr='$MYIP'/> </rule> <rule action='accept' direction='in' priority='500'> <udp dstipaddr='$MYIP'/> </rule> <rule action='accept' direction='in' priority='500'> <icmp dstipaddr='$MYIP'/> </rule> </filter> And the modifications made to the domain's network interface definition: <interface type='network'> <mac address='xx:xx:xx:xx'/> <source network='local'/> <target dev='vnet0'/> <filterref filter='my-static-ip'> <parameter name='MYIP' value=1.1.2.2'/> </filterref> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </interface> With the exception of simply manually adding iptables rules in place after I start the network using virsh, does anyone know how to accomplish what I'm trying to do? Ryan Sumner

1 0

[libvirt-users] schedinfo
by Zvi Dubitzky 16 Feb '11

16 Feb '11

libvirt schedinfo : ------------------ What is the meaning of the 'cap' parameter in the schedinfo virsh command (what are the units and range of values ) and what is the 'weigth' parameter ? Zvi Dubitzky Email:dubi@il.ibm.com

1 0

[libvirt-users] Clarification about virsh migration options
by Berend Dekens 16 Feb '11

16 Feb '11

I am trying to work out what all the options are for migrating a KVM machine to another KVM machine, without using shared storage. The documentation is not quite verbose and not intuitive, so I'm hoping someone can explain this to me. The man pages show this syntax: migrate optional --live --p2p --direct --tunnelled --persistent --undefinesource --suspend --copy-storage-all --copy-storage-inc domain-id desturi migrateuri dname The 'live' and 'suspend' options are clear. The 'undefinesource' option is straightforward as well. But what does 'persistent' mean? I mean, when transferring a VM to a destination, it will be available on the destination when migration completes, so what does 'persistent' mean in this context? The p2p, direct and tunneled options are not. When migrating, I assume it is possible to let the underlying virtualisation framework handle the migration - so I assume 'direct' means Qemu-KVM gets to migrate the VM and 'tunneled' means the RPC mechanism of libvirt migrates the machine. But what does 'p2p' mean? Normally, peer-to-peer implies direct communication, but since there is a 'direct' mode, I'm clueless what this option does. On a side note: is it possible to use tunneled mode to transfer VMs from for example Xen to KVM? Then I have some questions about the non-shared storage migration. I really like this option as most of my virtualized servers are run on one or two physical systems without shared VM storage. Migrating those VMs without downtime would be awesome. If I migrate the VMs, I assume the storage of the VM is places in the libvirt default storage location. But what is the difference between 'copy-storage-all' and 'copy-storage-inc'? Incremental hints that it would save incremental changes without actually transferring the storage between hosts. But how is that possible if the storage was not shared in the first place? Lots of questions but I hope someone who has done this before can answer some of these things. Regards, Berend Dekens

3 3

[libvirt-users] CfP 6th Workshop on Virtualization in High-Performance Cloud Computing (VHPC'11)
by VHPC2011 16 Feb '11

16 Feb '11

Apologies if you received multiple copies of this message. ================================================================= CALL FOR PAPERS 6th Workshop on Virtualization in High-Performance Cloud Computing VHPC'11 as part of Euro-Par 2011, Bordeaux, France ================================================================= Date: August 30, 2011 Euro-Par 2011: http://europar2011.bordeaux.inria.fr/ Workshop URL: http://vhpc.org SUBMISSION DEADLINE: Abstracts: May 2, 2011 Full Paper: June 13, 2011 Scope: Virtualization has become a common abstraction layer in modern data centers, enabling resource owners to manage complex infrastructure independently of their applications. Conjointly virtualization is becoming a driving technology for a manifold of industry grade IT services. The cloud concept includes the notion of a separation between resource owners and users, adding services such as hosted application frameworks and queuing. Utilizing the same infrastructure, clouds carry significant potential for use in high-performance scientific computing. The ability of clouds to provide for requests and releases of vast computing resource dynamically and close to the marginal cost of providing the services is unprecedented in the history of scientific and commercial computing. Distributed computing concepts that leverage federated resource access are popular within the grid community, but have not seen previously desired deployed levels so far. Also, many of the scientific datacenters have not adopted virtualization or cloud concepts yet. This workshop aims to bring together industrial providers with the scientific community in order to foster discussion, collaboration and mutual exchange of knowledge and experience. The workshop will be one day in length, composed of 20 min paper presentations, each followed by 10 min discussion sections. Presentations may be accompanied by interactive demonstrations. It concludes with a 30 min panel discussion by presenters. TOPICS Topics include, but are not limited to, the following subjects: - Virtualization in cloud, cluster and grid environments - VM-based cloud performance modeling - Workload characterizations for VM-based environments - Software as a Service (SaaS) - Cloud reliability, fault-tolerance, and security - Cloud, cluster and grid filesystems - QoS and and service levels - Cross-layer VM optimizations - Virtualized I/O and storage - Virtualization and HPC architectures including NUMA - System and process/bytecode VM convergence - Paravirtualized driver development - Research and education use cases - VM cloud, cluster distribution algorithms - MPI on virtual machines and clouds - Cloud frameworks and API sets - Checkpointing of large compute jobs - Cloud load balancing - Accelerator virtualization - Instrumentation interfaces and languages - Hardware support for virtualization - High-performance network virtualization - Auto-tuning of VMM and VM parameters - High-speed interconnects - Hypervisor extensions and tools for cluster and grid computing - VMMs/Hypervisors - Cloud use cases including optimizations - Performance modeling - Fault tolerant VM environments - VMM performance tuning on various load types - Cloud provisioning - Virtual machine monitor platforms - Pass-through VM device access - Management, deployment of VM-based environments PAPER SUBMISSION 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, accepted papers will be requested to provided source files. Format Guidelines: http://www.springer.de/comp/lncs/authors.html Submission Link: http://edas.info/newPaper.php?c=10155 CHAIR Michael Alexander (chair), IBM, Austria Gianluigi Zanetti (co-chair), CRS4, Italy PROGRAM COMMITTEE Paolo Anedda, CRS4, Italy Volker Buege, University of Karlsruhe, Germany Giovanni Busonera, CRS4, Italy Roberto Canonico, University of Napoli, Italy Tommaso Cucinotta, Scuola Superiore Sant'Anna, Italy William Gardner, University of Guelph, Canada Werner Fischer, Thomas-Krenn AG, Germany Wolfgang Gentzsch, Max Planck Gesellschaft, Germany Marcus Hardt, Forschungszentrum Karlsruhe, Germany Sverre Jarp, CERN, Switzerland Shantenu Sjha, Louisiana State University, USA Xuxian Jiang, NC State, USA Kenji Kaneda, Google, USA Simone Leo, CRS4, Italy Ignancio Llorente, Universidad Complutense de Madrid, Spain, Naoya Maruyama, Tokyo Institute of Technology, Japan Jean-Marc Menaud, Ecole des Mines de Nantes, France Anastassios Nanos, National Technical University of Athens, Greece Jose Renato Santos, HP Labs, USA Deepak Singh, Amazon Webservices, USA Boria Sotomayor, University of Chicago, USA Yoshio Turner, HP Labs, USA Kurt Tutschku, University of Vienna, Austria Lizhe Wang, Indiana University, USA Chao-Tung Yang, Tunghai University, China DURATION: Workshop Duration is one day. GENERAL INFORMATION The workshop will be held as part of Euro-Par 2011, organized by INRIA, CNRS and the University of Bordeaux I, II, France. Euro-Par 2011: http://europar2011.bordeaux.inria.fr/

1 0

[libvirt-users] Use libvirt-python to attach cdrom
by seven 16 Feb '11

16 Feb '11

Hi all, I want to attach cdrom by libvirt.py .But can't find a func in libvirt.py can finish this work. Who can help me ? Regard -- Seven Ling E-mail: jiejie.ling(a)qq.com

2 3

[libvirt-users] monitiring cpu usage via cgroup
by Zvi Dubitzky 16 Feb '11

16 Feb '11

thanks Kame, 1) the policy of my process(test) is indeed SCHED_NORMAL . When I use sched_setscheduler() to set the pid sched policy to SCHED_FIFO then the cpu usage via cgroup rt_runtime_us and rt_period_us is affected . Still the ratio I specified is 1:10 but I see via top that my process ,that gets 1 cpu ,shows 20% usage 2) What about memory and cpuset subsystems of cgroup. Do their settings apply also only in the cases of processes with policy = SCHED_FIFO/SCHED_RR ? 3) lastly: What is the meaning of the ‘cap’ parameter in the schedinfo virsh command (what are the units and range of values ) and what is the ‘weigth’ parameter .I did not find any documentation of libvirt about it ? thanks Zvi Dubitzky Email:dubi@il.ibm.com

1 0

[libvirt-users] monitiring cpu usage via cgroup
by Zvi Dubitzky 15 Feb '11

15 Feb '11

Hi I was asking about the fedora 14 kernel if it is good enough for cgroup usage because I am trying to set a cgroup under cpu subsytem ( /dev/cgroup/cpu/group1/ ) that have /cpu.rt_runtime_us of 100000 while cpu.rt_period_us has a value of 1000000 i.e a ratio of 1/10 . still when I run a task (endless loop) in that group (cgexec -g cpu,cpuset:group1 ./test) it gets all the cpu core time that is assigned to it ( i watch via top utility) , so it seems that the quota set via the group does not take effect although I restart the cgconfig service . I also verified libcgroup is installed by: rpm -q libcgroup Am I missing something or top is not the right utility to watch the cpu usage in this case ? thanks Zvi Dubitzky Email:dubi@il.ibm.com

2 1

[libvirt-users] Xen disk device detach fails as non-root [libvirt-0.8.7 and older versions]
by Iain MacDonnell 15 Feb '11

15 Feb '11

Hi All, I find that I am able to attach a disk device do a Xen domain, using virDomainAttachDevice(), running as a non-root user, but I am unable to use virDomainDetachDevice() - it results in an "unknown failure". Using "virsh [attach|detach]-device" exhibits this behviour. $ virsh attach-device domu1 attach.xml Device attached successfully $ virsh detach-device domu1 attach.xml error: Failed to detach device from attach.xml error: Unknown failure $ With some digging, I determined that the problem arises when libvirt tries to translate the device name to a number, using the XenStore API (xenStoreDomainGetDiskID()), which requires use of the "xenstored" UNIX socket, and that socket is only accessible by root. On making that socket accessible to the user (by group), virDomainAttachDevice() starts working, but I'm then unable to list domains, because xenStoreDoListDomains() waits to verify each domain using xenHypervisorHasDomain(), and that requires access to another socket - "/proc/xen/privcmd" My question, before going down the path of trying to hack permissions for these sockets permanently ..... is this how it's supposed to be, or could, perhaps, libvirtd, which runs as root, access these sockets on behalf of the user? It seems it should at least fail more gracefully.... TIA for any pointers.... ~iain

2 1

[libvirt-users] connecting to virtualbox with libvirt
by Gary Scarborough 14 Feb '11

14 Feb '11

I am trying to connect to the virtualbox hypervisor through libvirt. If I am doing it correctly, virsh -c vbox:///session should get me connected. I am on F14 64 bit. I tried this with both vbox 3.2 and 4.02. I keep getting the following error: error: no connection driver available for vbox:///session. What am I missing? Thanks, Gary

2 1