- Users - libvirt List Archives

Re: [libvirt-users] [libvirt] configuring network interface
by Marcela Castro León 11 Feb '11

11 Feb '11

Hello Laine: I've revised the configuration and followed all your instructions. I'll try to give you a better explain of my problem. a) I need a bridged (3) configuration, because I have to connect the guests from each others. (via ssh and execute mpi). I've three hosts, each of one has at least a guest. I'll only explain about the configuration of one of them, because the problem is only related a one configuration. hostname=rionegro ip=192.168.1.4 guest-name=rnompi1 b) I've configured host bridge following step by step ther instruction on http://wiki.libvirt.org/page/Networking this is the /etc/network/config # The loopback network interface auto lo iface lo inet loopback auto br0 iface br0 inet static address 192.168.1.4 netmask 255.255.255.0 network 192.168.1.0 broadcast 192.168.1.255 gateway 192.168.1.10 bridge_ports eth0 bridge_stp on bridge_fd 0 bridge_maxwait 0 c) this is the xml guest configuration related to the network. <interface type='bridge'> <source bridge='br0'/> </interface> d) But the guest doesn't get any IP automatically, and create an interface with a variable name (eth4, eth5,etch6..)*. I don't know how to achieve the guest don't behave like this and create always its network interface *with a fixed name, for example, eth0). Why I ask that? if the guest created an interface with a fixed name, I would configure and static ip for the it. I'm attaching a printscreen "rnompi1.jpg" that show what ifconfig ahow on the guest (eth14 and without any IP). It doesn't get their host either. The only way I achieve to solve but "temporally" until reboot is configuring the guest statically using the name of the interface that has just up.In that case every works fine, and the guest can connect to all the other guests on the lan and others can connect to it. But this only works until reboot or reconfigure the guest, because each time the name of interface change, and my configuration I had done of /etc/network/interface didn't work any more because is related to another interface. I hope I've explained better the problem I have. Thank you very much, you've already helped too much to better understand the different cases of network configuration Regards. 2011/2/10 Marcela Castro León <mcastrol(a)gmail.com> > Hello Laine: > Thank you very much. > I've showed only one of the guest, each one have it's own ip address. > Anyway, I'll follow your advise of revise all the configuration after > reading the doc. > Thank you very much again. > Regards. > > 2011/2/8 Laine Stump <laine(a)laine.org> > > (For future questions such as this, you may find you get a better response >> (and lower signal-noise ratio) by asking on the recently -created >> libvirt-users(a)redhat.com list, as libvir-list has a very large volume of >> patches and discussion about the internals of libvirt) >> >> >> On 02/08/2011 12:05 PM, Marcela Castro León wrote: >> >> Hello >> >>> >>> I've a lan of virtual machine configured with bridge. I've a private >> lan of 4 machine, but one of them have 2 lan-card with access to internet. >> >> Evertything works almost fine, except that I've configured on each guest >> the static configuration of the bridge with the ip address, like this: >> >> auto lo >> iface lo inet loopback >> >> auto eth6 >> iface eth6 inet static >> address 192.168.1.41 >> network 192.168.1.0 >> netmask 255.255.255.0 >> gateway 192.168.1.10 >> >> >> >> So are you saying that you're attempting to give each guest the *same* IP >> address? There's no way that can possibly work, even if each is connected to >> a separate virtual network or bridge. Since the host's IP stack can "see" >> all of the guests' network traffic, regardless of the connection mode you >> choose, each must have a unique IP address even if they can't see each >> other. >> >> >> >> >> But each time I, for example, define the guest, it start with a ascending >> number of interface: eth6,eth7, eth8, etc, Now is initializing with eth12. >> >> >> There must be a mixup in the explanation - I'm understanding that you put >> the same config file in each guest, but each guest, when it starts, shows a >> different ethernet interface in its ifconfig. That doesn't seem right, >> because none of the guests know about each other. Or are you saying >> something different? Where are you seeing the "eth6", "eth7", ... "eth12"? >> In the guest XML config maybe? >> >> >> So, it doesn't get the manual configuration I've done on the >> /etc/network/interfaces. >> How can avoid this way of work of the number of ethernet interface the >> guest start? >> >> >> Why are you doing a manual/static definition on the guest at all? It's >> simpler to just leave the guest to automatically configure its eth0, and if >> you really want a particular IP address for a particular guest, handle that >> with a <host> element in the <dhcp> section of the virtual network the guest >> is connected to (or, if you're using a pure bridging setup, then in the >> static hosts config of the dhcp server on that network). >> >> >> >> A way i'm trying to solve is configuring the interface trough the guest >> xml definition, >> >> >> The guest xml interface configuration cannot be used to setup the IP >> address, route, etc config of the guest. It is only used for two things: >> >> 1) configuring what kind of hardware appears on the guest (but not what >> name the guest gives to it!) >> >> 2) configuring how that hardware is connected to the physical network >> (usually either via an existing bridge interface on the host (in order for >> the guest to appear as if its directly connected to the physical network), >> or via a "virtual network" created by libvirt (which is really just a bridge >> that's not directly connected to any physical interface, relying instead on >> the host's IP stack to route packets from the guest out to the physical >> network.) >> >> >> but the operation "define" on virsh accepted without error the xml i've >> done, but the dumpxml doesn't show them, so, it doesn't work at all. >> I've followed the instruction that said the manual "Application >> development guide - a guide to application development with libvirt" on page >> 71. >> >> >> >> That document is targeted towards software developers writing applications >> to manage virtual guests using the libvirt API, not for system >> administrators who are configuring virtual guests manually (or, more likely, >> by using software that uses the libvirt API, eg virsh or virt-manager). It >> likely contains more detail and less overview than what you need, which is >> probably what led to the confusion. I've included a pointer to a wiki page >> below that is more the appropriate level of information (Justin, if the docs >> you were working on are somewhere easily accessible, you can point to those >> instead.) >> >> >> >> The whole definition of the VM is attached, but the related to interface >> is: >> <interface type='bridge'> >> <source bridge='br0'/> >> <target dev='vnet0'/> >> <protocol> >> <ip address="192.168.1.41" prefix="24"/> >> <route gateway="192.168.1.10"/> >> </protocol> >> <bridge stp="off" delay="0.01"> >> <interface type="ethernet" name="eth12"> >> <protocol> >> <ip address="192.168.1.41" prefix="24"/> >> <route gateway="192.168.1.10"/> >> </protocol> >> </interface> >> </bridge> >> </interface> >> >> >> >> You are confusing the guest interface config (that is part of the guest's >> XML config, and described above), and host physical interface config (that >> is not in any way associated with guest config, but instead used to >> configure the host's physical network interfaces, bridges, bonds, and vlan >> interfaces; and nearly as important, is not supported on Ubuntu anyway). >> >> Actual guest interface config is *much* simpler than that. >> >> >> >> >> But after define, the dumpxml only show this part: >> <interface type='bridge'> >> <mac address='52:54:00:3c:92:9a'/> >> <source bridge='br0'/> >> </interface> >> >> >> >> That's because everything else you've put in your guest config is not >> actually a part of a valid guest interface config, and so is ignored. >> >> >> >> >> I would appreciate any help to solve the question. The problem I've is I >> lost the access to the guest every time the libvirt change the name of the >> interface eth1,2,3, etc. >> >> >> In this case I think it may be useful to start over. You've somehow been >> tricked into misunderstanding the way that network configuration in libvirt >> works, confusing host interface config with guest interface config, and the >> resulting config bears little resemblance to what would actually work. >> >> == >> The first step is to read this page: >> >> http://wiki.libvirt.org/page/Networking >> >> to get a basic overview. >> >> == >> Second step: you should decide which of the three methods of connecting to >> guests to the physical network you want to use: >> >> 1) NATed >> 2) routed >> 3) bridged >> >> The simplest is (1), and that will probably work just fine as long as 1) >> you don't need incoming connections to the guests, and 2) you're not >> concerned about getting the last fraction of an ounce of performance out of >> the network. >> >> Depending on which of those 3 you decide to use: >> >> == >> Third step: follow the instructions to either create a bridge interface on >> the host connected to a physical network interface (for (3)), or to create a >> libvirt virtual network using one of the two modes ((1) or (2) - actually >> the stock libvirt config already has a "default" network that is probably >> adequate for (1), and unless you understand IP routing very well, you'll >> likely want to avoid (2)). >> >> Once the physical network is setup: >> >> == >> Fourth step: define your guests. In each guest, you'll have an <interface> >> element, but it will be much simpler than what you've previously tried. For >> bridged mode, the interface section would look like this: >> >> >> <interface type='bridge'> >> <source bridge='br0'/> >> </interface> >> >> Depending on the guest, you may want to specify what type of hardware is >> presented to the guest (default is rtl, which almost always works, but is >> not the best perfrorming), in particular, if the guest supports virtio-net, >> you'll want to add this into the <interface> element: >> >> <model type='virtio'/> >> >> For NATed or routed mode, the guest's interface definition would look like >> this: >> >> <interface type='network'> >> <source network='default'/> >> </interface> >> >> (replace "default" with the name of the network you've defined, if you've >> defined your own). In this case also, you can add a <model type='...'/> if >> appropriate. >> >> Note that I don't put anything in the <interface> element about MAC >> address, target device, or bus address. Those are all setup automatically by >> libvirt the first time you start the guest, and it will make sure that the >> guests don't interfere with each other. >> >> == >> Fifth step: only if you need a particular guest to have a particular fixed >> IP address, either 1) edit the network config file on the guest, 2) add a >> static host to the libvirt virtual network definition (giving the MAC >> address provided to the guest by libvirt during the initial guest startup) >> with a <host> element added to the <dhcp> element of the network; details >> are here: >> >> http://www.libvirt.org/formatnetwork.html >> >> or 3) if you're using the bridged method to connect to the physical >> network, again either configure the guest's own config file, or add a static >> host definition to the dhcp server running on your physical network. >> >> If you still have problems don't hesitate to send another mail (although, >> as I said at the top, sending to libvirt-users(a)redhat.com may yield >> better results than libvir-list(a)redhat.com) >> > >

3 5

[libvirt-users] How to retrieve Memory stat of the KVM using python binding with Libvirt API
by arpita k 10 Feb '11

10 Feb '11

Hi , I am using ubuntu host machine and KVM is present as guest in that machine.Two virtual machines are there.I am using Python binding to query on the hypervisor and extract the VM related information.I like to print memory stat of the VM like SWAP_IN , SWAP_OUT,free memory, mem cached etc.So, the code, I have written for this, is : #!/usr/local/bin/python import libvirt import sys conn = libvirt.openReadOnly(None) if conn == None: print "Failed to open connection to the hypervisor" sys.exit(1) try: domIds = conn.listDomainsID() for id in domIds: dom0 = conn.lookupByID(id) try: print "memory stats: %s"% dom0.memoryStats().values() except: print 'Failed to find the memory data of this domain' sys.exit(1) except: print 'Failed to find the main domain' sys.exit(1) But this code 'dom0.memoryStats().values()' is returning empty list.I need to collect more information regarding the memory .How should I proceed? Regards, Arpita

1 0

Re: [libvirt-users] [virt-tools-list] store vm xml files in another directory
by Cole Robinson 10 Feb '11

10 Feb '11

On 02/08/2011 03:05 PM, Kenneth Armstrong wrote: > I know that by default, vm xml files are in > /etc/libvirt/qemu/vm-name.xml, but is there a configuration somewhere > that I could point to another directory? For instance, have it on > something like /srv/qemu/xml/vm-name.xml which would be a GFS2 > filesystem (for HA purposes)? > It isn't configurable, short of modifying the source code and recompiling. As far as sharing /etc/libvirt/qemu, Dan has responded to that question in the past: http://www.mail-archive.com/libvir-list@redhat.com/msg16921.html - Cole

3 4

[libvirt-users] How to use Python Binding with Lbivirt.
by arpita k 09 Feb '11

09 Feb '11

HI, I am using ubuntu host machine and KVM is present as guest in that machine.Two virtual machines are there.I am using Python binding to query on the hypervisor and extract the VM related information. Here is my simple code: #!/*usr*/local/bin/*python * import libvirt import sys conn = libvirt.openReadOnly(None) if conn == None: print *'Failed to open connection to the hypervisor' * sys.exit(1) try: domIds = conn.listDomainsID() for id in domIds: dom0 = conn.lookupByID(id) dom0.info() print *"Domain 0: id %d running %s"* % (dom0.ID(), dom0.OSType()) print dom0.info() except: print *'Failed to find the main domain' * sys.exit() The output is : Domain 0: id 2 running hvm [1, 131072L, 131072L, 1, 14921710000000L] Domain 0: id 3 running hvm [1, 131072L, 131072L, 1, 30880000000L] Here I am not able to understand these values are represent what type of information, which value is representing what resource of host machine.Could you please help me to understand how I could find out CPU,VM ,memory related information seperately through python binding?. I didn't find Python binding Reference in the Libvirt site.Could you help me by providing the Reference link for Python binding? Thanks in Advance!! Arpita

2 1

[libvirt-users] Making multiple connections from an application using libvirt-0.6.3-20.1.el5_4
by Nandini Chandra 09 Feb '11

09 Feb '11

I came across a thread in the archives that seems to suggest that multiple connections can be made from a single application,although I am unsure what version(s) of libvirt support it. https://www.redhat.com/archives/libvirt-users/2010-March/msg00074.html I'd like to know if multiple connections(for eg,virConnectOpenAuth()) can be made from an application using libvirt-0.6.3-20.1.el5_4 . Thanks

3 2

[libvirt-users] Question about putting vm in blkio cgroup
by Nikita A Menkovich 09 Feb '11

09 Feb '11

Hello, I have another question about cgroup in libvirt I create a script that put vm in cgroup for limiting blkio sudo cgcreate -f 750 -d 750 -a $user:$group -g blkio:$alias pid=`pgrep $alias` echo $pid > /dev/cgroup/blkio/$alias/tasks $virsh schedinfo --set cpu_shares=$cpu $alias echo "8:0 $disk_io" > /dev/cgroup/blkio/$alias/blkio.throttle.read_bps_device echo "8:0 $disk_io" > /dev/cgroup/blkio/$alias/blkio.throttle.write_bps_device For example I write: 8:0 10485760 to /dev/cgroup/blkio/$alias/blkio.throttle.*_bps_device So there should be limit to 10Mbyte/s to acces to /dev/sda But when I run hdparm -t /dev/vda or dd to test limits I receive value ~ 70 Mbyte/s -- Nikita A Menkovich JID: menkovich(a)gmail.com

1 0

[libvirt-users] Serving DHCP and DNS from a KVM guest
by Adam Reid 09 Feb '11

09 Feb '11

Hello All, I am trying to set up a cobbler server running as a guest on an isolated network. I'm having trouble figuring out how I can forward the DHCP and DNS traffic to this guest instead of DNSMASQ on the host. So far I have only tried to get dhcp working by enabling the 'allow-dhcp-server' on a guest named test and setting the $DHCPSERVER variable through a filterref parameter element. I figure If I can get this working I can use this filter as a basis for DNS. In my guest the filter is applied to the only network interface setting DHCPSERVER to the IP of my cobbler VM. <interface type='network'> <mac address='52:54:00:6b:56:ab'/> <source network='internal'/> <target dev='vnet2'/> <model type='virtio'/> <filterref filter='allow-dhcp-server'> <parameter name='DHCPSERVER' value='192.168.100.2'/> </filterref> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> And the internal network is defined as: <network> <name>internal</name> <uuid>c1aef8ea-9d30-7514-5eb4-5ffbc235cffb</uuid> <bridge name='virbr1' stp='on' delay='0' /> <ip address='192.168.100.1' netmask='255.255.255.0'> </ip> </network> My understanding of the filter is that when it is applied to a hosts interface it should allow DHCP requests to go to a server other than the host system. Is this correct? Are guests allowed to serve DHCP to other guests? Does anyone have any suggestions or experience with this? Thanks in advance, Adam Reid

1 0

[libvirt-users] Libvirt in Windows not working
by arpita k 09 Feb '11

09 Feb '11

Hi, This is Arpita. I am using Libvirt API in the Windows Vista system.VMware is installed over there.I have installed Libvirt and Python 2.6 in the Windows Vista system andhave written a small test file in Python. Here is the code: import libvirt* * import sys conn = libvirt.openReadOnly(None) if conn == None: print *'Failed to open connection to the hypervisor' * sys.exit(1) try: dom0 = conn.lookupByName(*"vmware1"*) except: print *'Failed to find the main domain' * sys.exit(1) print *"Domain 0: id %d running %s"* % (dom0.ID(), dom0.OSType()) print dom0.info() When I am trying to execute this python code I received error message that : ImportError: No module named libvirt Please help me. Regards, Arpita

3 2

[libvirt-users] Disk atomicity...
by Dale Amon 07 Feb '11

07 Feb '11

I suspect the answer is no, as the very idea makes my head hurt, but a confirmation would be nice: Am I correct in assuming that it is a very bad idea to mount a physical disk on both the physical machine and on a hosted VM at the same time? I can imagine getting away with it if one were RO, but not otherwise. Any takers?

3 2

[libvirt-users] Live migration using virsh
by rajalakshmi 06 Feb '11

06 Feb '11

Sir, I am working in cloud environment and i have set up a private cloud using eucalyptus. I am trying to migrate virtual machines within the cloud using virsh migrate command. I could successfully implement it but when i wanted to know the method used to do live migration. I tried to interpret the code. I also saw that they are setting a flag initially if it is live migration and then assigning 1 to variable live. and i also could find that xenDomainMigratePrepare is calling xend_op which in turn is calling xend_op_ext. But i could'n understand the method of migration. Can you please explain the logic?

1 0

[libvirt-users] Network/bridge questions
by Alex 06 Feb '11

06 Feb '11

Hi, I have an fc14 install and would like to install a few kvm guests but am having difficulty with the networking. On the host I have disabled NetworkManager and configured a bridge which has eth0, the only physical interface on the server. I have a dhcp server on the local lan, and if I add a mac address entry to my dhcpd.conf, the guest will find an IP from there. I think somehow the guest interface is not properly being routed through the virtual gateway created by libvirtd, and is instead just routed through br0 directly on the host. What am I missing? Here is some relevant output from some of the things I know: # # brctl show bridge name bridge id STP enabled interfaces br0 8000.485b3961b55c no eth0 vnet0 virbr0 8000.000000000000 yes I don't understand what purpose the vnet0 has, or even how it got there in the first place. Is it necessary, or did I do something wrong here? # virsh net-list --all Name State Autostart ----------------------------------------- default active yes >From the guest xml file in /etc/libvirt/qemu/ <interface type='bridge'> <mac address='52:54:00:a8:c5:9d'/> <source bridge='br0'/> <model type='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/> </interface> I've also flushed the firewall rules, so I'm sure the dhcp request isn't being blocked. I've also tried changing the br0 above to virbr0, with no change. This is from "net-edit default" in virsh: <network> <name>default</name> <uuid>3d1f153f-fc0f-4de5-98c8-7f99f96f34d6</uuid> <forward mode='nat'/> <bridge name='virbr0' stp='on' delay='0' /> <ip address='192.168.122.1' netmask='255.255.255.0'> <dhcp> <range start='192.168.122.2' end='192.168.122.254' /> </dhcp> </ip> </network> Sure would appreciate any ideas you may have. Thanks, Alex

2 1

[libvirt-users] libvirt, snapshots, LVM, and existing VMs
by Casey Link 05 Feb '11

05 Feb '11

Hello, I am running libvirt 0.8.3 with a couple VMs using a LVM VG that I had previously setup on my host. I did not use the storage pool functionality of libvirt to create the VG in the first place ('pool-list' is empty). Now I want to be able to take snapshots of the VMs. Having read this previous post [1] on the mailing list, I gather that libvirt does indeed support taking snapshots of VMs on an LVM. I suppose I need to create storage pool that includes the <backingStore> element. However, I am not sure how to go about creating this pool and making my existing VMs utilize it. Is what I am trying to do even possible? If so, what is the proper way to go about doing it? Best, Casey Link [1] http://www.redhat.com/archives/libvir-list/2009-February/msg00196.html

3 7

[libvirt-users] Release of libvirt-java 0.4.7
by Bryan Kearney 04 Feb '11

04 Feb '11

I have just released 0.4.7 of libvirt java. There are 2 main items in this release: - Only throw errors on real errors. - Remote non thread safe error reporting - BZ 600819 Incorrect scheduler parameter value passed to native API. You can access the latest version via the following means: Source Code: http://www.libvirt.org/git/?p=libvirt-java.git;a=summary Bundled Source (tarball and SRPM): http://libvirt.org/sources/java/ Maven: http://libvirt.org/maven2/ RPMS are making their way through F-14, F-13, and F12 build systems Thank you! -- bk

1 0

[libvirt-users] iSCSI storage pool questions
by Anthony Goddard 03 Feb '11

03 Feb '11

Hi All, I've been trying to figure out the best way of using an iSCSI SAN with KVM and thanks to a helpful post by Tom Georgoulias that I found on this list (https://www.redhat.com/archives/libvirt-users/2010-May/msg00008.html) it appears I have a solution. What I'm wondering is the following: 1) If I use an iSCSI LUN as the storage pool (instead of creating an LVM VG from this iSCSI LUN manually first and using LVM in libvirt) do I loose any functionality (such as snapshotting)? 2) If I am intending on doing offline & live-migration between hosts, does each host have to connect to the same LUN / storage pool, or should each host have a separate LUN that is stores VMs on and then each host should simply see all of the LUNs so that migration is possible? For example: I create a LUN on the SAN, connect this to host 1 as a pool called "pool 1" and create VMs on it. Then, on host 2, do I connect to the same "pool 1" and create VMs there, or do I just connect to "pool 1" in order to see it (for migrations) but also define "pool 2" from another new LUN on the SAN and create VMs there? Hope that makes sense.. I'm trying to understand in what situations I might get myself in trouble by two hosts performing storage operations at the same time on the same pool (which is unlikely, but I'm trying to figure out where I stand) Thanks! Anthony

1 1

[libvirt-users] virsh : Need help to manage a vm on a remote vmWare
by Etienne GOSSET 03 Feb '11

03 Feb '11

Hello, I'm trying to use the tool virsh of libvirt. I want to manage via CLI a vm on a vmware hypervisor which is hosted on a remote host (Is that what you call the node??). I read a lot of thing on your site but I really don't understand how to do something... I saw that: >virsh edit $your-vm-name I think "$your-vm-name" is the xml file you're talking about? Is it mandatory to use those XML file in order to use virsh? (Because it is very huge!) I also saw this way of using virsh : >virsh --connect esx://user@hypervisorHost://path/to/myVM.vmx Result is: error: invalid argument in libvirt was built without the 'esx' driver What is the problem here? For information, I installed libvirt-0.8.7, can you give me the first step to manage an existing vm hosted on a distant vmWare (via CLI). Thank you in advance.

2 1

[libvirt-users] network-performance
by Orion Poplawski 02 Feb '11

02 Feb '11

I'm just starting to take a look at guest networking performance and am a little disappointed. I'm comparing two setups: Host: Windows Server 2008 R2 Hyper-V Guest: CentOS 5.5 x86_64 Host: CentOS 5.5 x86_64 kvm running libvirt Guest: CentOS 5.5 x86_64 The guests are essentially identical except that I'm running the Microsoft Linux Integration Components synthetic drivers on the windows hosted VM. The libvirt setup uses bridged networking. Running bonnie++ on a nfs mounted filesystem on each guest I'm seeing the libvirt hosted guest get between 16%-35% of the performance of the Hyper-V guest. Is this expected? Is there anything I can do to increase network performance of the kvm guest? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion(a)cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com

5 6

[libvirt-users] Not getting a virsh console
by Nikunj A. Dadhania 02 Feb '11

02 Feb '11

Hi, I am seeing a problem with a earlier working domain configuration. Here is the domain xml file. I am not getting the virsh console now. I am trying this on RHEL 6.0. Here is the list of steps that I am using: [root@krm1 test]# virsh define vmtest.xml Domain vmtest3 defined from vmtest.xml [root@krm1 test]# virsh start vmtest3 Domain vmtest3 started [root@krm1 test]# virsh console vmtest3 Connected to domain vmtest3 Escape character is ^] [root@krm1 test]# virsh dominfo vmtest3 Id: 2 Name: vmtest3 UUID: 2b38ff93-1bc6-c78c-1ed8-ffb936c86dac OS Type: hvm State: running CPU(s): 1 CPU time: 17.3s Max memory: 524288 kB Used memory: 524288 kB Persistent: yes Autostart: disable [root@krm1 test]# The state of the domain shows running. One more thing that i noted was that qemu-kvm is consuming 100% of the cpu. So not sure if the qemu-kvm has started properly. I am not sure how to debug the issue. Would appreciate some help on this. Some more information: I have attached following file /var/log/libvirt/qemu/vmtest3.log Executing the following command on the bash works fine and the VM comes up: /usr/libexec/qemu-kvm -M rhel6.0.0 -enable-kvm -m 512 -smp 1,sockets=1,cores=1,threads=1 -drive file=/home/bharata/cgroup-scalability/test/test.qcow2 -nographic While executing the command in the vmtest3.log, the VM does not come at all. Here is the domain xml: <domain type='kvm'> <name>vmtest3</name> <memory>524288</memory> <vcpu>1</vcpu> <os> <type arch='i686'>hvm</type> <boot dev='hd'/> </os> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/libexec/qemu-kvm</emulator> <disk type='file' device='disk'> <driver name='qemu' type='qcow2'/> <source file='/home/bharata/cgroup-scalability/test/test.qcow2'/> <target dev='hda' bus='ide'/> </disk> <serial type='pty'> <target port='0'/> </serial> <console type='pty'> <target port='0'/> </console> </devices> </domain> Thanks, Nikunj LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin /usr/libexec/qemu-kvm -S -M rhel6.0.0 -cpu qemu32 -enable-kvm -m 512 -smp 1,sockets=1,cores=1,threads=1 -name vmtest3 -uuid 2b38ff93-1bc6-c78c-1ed8-ffb936c86dac -nographic -nodefconfig -nodefaults -chardev socket,id=monitor,path=/var/lib/libvirt/qemu/vmtest3.monitor,server,nowait -mon chardev=monitor,mode=control -rtc base=utc -no-acpi -boot c -drive file=/home/bharata/cgroup-scalability/test/test.qcow2,if=none,id=drive-ide0-0-0,boot=on,format=qcow2 -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -chardev pty,id=serial0 -device isa-serial,chardev=serial0 -usb -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x2 09:53:22.536: debug : virCgroupNew:493 : New group /libvirt/qemu/vmtest3 09:53:22.536: debug : virCgroupDetect:230 : Detected mount/mapping 0:cpu at /cgroup/cpu in 09:53:22.536: debug : virCgroupDetect:230 : Detected mount/mapping 1:cpuacct at /cgroup/cpuacct in 09:53:22.536: debug : virCgroupDetect:230 : Detected mount/mapping 2:cpuset at /cgroup/cpuset in 09:53:22.536: debug : virCgroupDetect:230 : Detected mount/mapping 3:memory at /cgroup/memory in 09:53:22.536: debug : virCgroupDetect:230 : Detected mount/mapping 4:devices at /cgroup/devices in 09:53:22.536: debug : virCgroupDetect:230 : Detected mount/mapping 5:freezer at /cgroup/freezer in 09:53:22.536: debug : virCgroupMakeGroup:451 : Make group /libvirt/qemu/vmtest3 09:53:22.536: debug : virCgroupMakeGroup:463 : Make controller /cgroup/cpu/libvirt/qemu/vmtest3/ 09:53:22.536: debug : virCgroupMakeGroup:463 : Make controller /cgroup/cpuacct/libvirt/qemu/vmtest3/ 09:53:22.536: debug : virCgroupMakeGroup:463 : Make controller /cgroup/cpuset/libvirt/qemu/vmtest3/ 09:53:22.536: debug : virCgroupMakeGroup:463 : Make controller /cgroup/memory/libvirt/qemu/vmtest3/ 09:53:22.536: debug : virCgroupMakeGroup:463 : Make controller /cgroup/devices/libvirt/qemu/vmtest3/ 09:53:22.536: debug : virCgroupMakeGroup:463 : Make controller /cgroup/freezer/libvirt/qemu/vmtest3/ 09:53:22.536: debug : virCgroupSetValueStr:275 : Set value '/cgroup/cpu/libvirt/qemu/vmtest3/tasks' to '7312' 09:53:22.545: debug : virCgroupSetValueStr:275 : Set value '/cgroup/cpuacct/libvirt/qemu/vmtest3/tasks' to '7312' 09:53:22.554: debug : virCgroupSetValueStr:275 : Set value '/cgroup/cpuset/libvirt/qemu/vmtest3/tasks' to '7312' 09:53:22.562: debug : virCgroupSetValueStr:275 : Set value '/cgroup/memory/libvirt/qemu/vmtest3/tasks' to '7312' 09:53:22.569: debug : virCgroupSetValueStr:275 : Set value '/cgroup/devices/libvirt/qemu/vmtest3/tasks' to '7312' 09:53:22.577: debug : virCgroupSetValueStr:275 : Set value '/cgroup/freezer/libvirt/qemu/vmtest3/tasks' to '7312' 09:53:22.585: debug : qemudInitCpuAffinity:2618 : Setting CPU affinity 09:53:22.586: debug : qemuSecurityDACSetProcessLabel:547 : Dropping privileges of VM to 107:107 char device redirected to /dev/pts/2

1 0

[libvirt-users] (no subject)
by David M. Barlieb 02 Feb '11

02 Feb '11

Hello all, I am looking to use bonded interfaces as a bridge device for my virtual hosts. My goal was to have 4 nics bonded together to create a large pipe for my guests. I then tried to create a bridge device out of it similar to creating a bridge out of a single nic. It unfortunately did not work. Does any one know of a way to bond 2 or more nics than configure the bond as a bridge to the virtual guests. Thanks in advance, Dave B

1 0

[libvirt-users] libvirt bridge configuration
by Taras Lipatov 01 Feb '11

01 Feb '11

Hi Every one, I'm hoping some one can point me in the right direction. We are piloting KVM + Libvirt for migrating off Xen and I'm stuck with the network bridge configuration via libvirt. What we have currently with xen is: 1 server with Wan/lan interfaces and a pool of public ip addresses. Using xen bridge scripts I'd start a xenbr0 and attach eth0 (wan) and go ahead and assign public IP's to images attached to xenbr0. Now with creating virbr0, I need to assign it an ip address? and routing methods? what about creating a transparent bridge via libvirt? I kno I can achieve this using bridge utils, but it would be nice to have everything handled by libvirt. Any help would be appreciated. cheers, Taras

4 4

[libvirt-users] Guest interface info under ESX driver
by Sengor 01 Feb '11

01 Feb '11

Hi, Is there a way to fetch the current IP address of a running guest domain (with VMware Tools installed) running on an ESX host? Ideally this would be doable via the libvirt Python bindings. Thank you. -- sengork

3 4

[libvirt-users] syslog messages after cd-rom removal
by Alexander Topolanek 31 Jan '11

31 Jan '11

Sorry, I know this is not really libvirt related, but I'm searching for a solution sind about an hour and cannot find anything that helps me ... Ok, I used to install a standard debian in a kvm instance, and removed the virtual cd-rom after installation. Now I'm getting this triplet every five minutes in my syslog: Jan 26 19:25:12 srv62 kernel: [ 1324.267675] hdc: task_in_intr: status=0x41 { DriveReady Error } Jan 26 19:25:12 srv62 kernel: [ 1324.267683] hdc: task_in_intr: error=0x04 { AbortedCommand } Jan 26 19:25:12 srv62 kernel: [ 1324.267685] ide: failed opcode was: 0xec and I'm looking how to get rid of the device. thanks for any help Alexander

3 4

[libvirt-users] compilation error
by antoni artigues 28 Jan '11

28 Jan '11

Hello I'm trying to compile libvirt-0.8.6 on RHEL4 with kernel 2.6.9-67 The make command returns: -------------------- util/interface.c: In function `ifaceGetVlanID': util/interface.c:355: error: `GET_VLAN_VID_CMD' undeclared (first use in this function) util/interface.c:355: error: (Each undeclared identifier is reported only once util/interface.c:355: error: for each function it appears in.) make[3]: *** [libvirt_util_la-interface.lo] Error 1 make[3]: Leaving directory `/root/InstalledPackages/libvirt-0.8.6/src' make[2]: *** [all] Error 2 make[2]: Leaving directory `/root/InstalledPackages/libvirt-0.8.6/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/root/InstalledPackages/libvirt-0.8.6' make: *** [all] Error 2 -------------------- I have saw that GET_VLAN_VID_CMD should be in if_vlan.h file, but it is not present. What do I have to update in my system to avoid this error? Thanks in advance regards Antoni Artigues

4 16

[libvirt-users] libvirtd + vir-manager + kerberos
by Thomas Schweikle 28 Jan '11

28 Jan '11

Hi! Having two hosts installed with libvirtd, kvm, qemu on (Ubuntu 10.10). Now I have one big problem and one less: I have set up kerberos for both hosts. Created the principal "libvirt/srv1.example.org(a)EXAMPLE.ORG" and "libvirt/srv2.example.org(a)EXAMPLE.ORG", Exported the krb5.keytab, Installed it and tested the servers: srv1.example.org: I can connect using kerberos after acquiring a ticket with kinit. srv2.example.org: I am asked for user and password. Setup seems to be identical. Is there a way to debug, what is going on on this server? I'd like to have both respect kerberos and allow logging in with no password it already authenticated! Here is what I've set up: /etc/sasl2/libvirt.conf listen_tls = 0 listen_tcp = 1 mdns_adv = 0 auth_unix_ro = "none" auth_unix_rw = "none" auth_tcp = "sasl" /etc/sasl2/libvirt.conf mech_list: gssapi keytab: /etc/libvirt/krb5.kqemu sasldb_path: /etc/libvirt/passwd.db I start libvirtd with: KRB5_KTNAME=/etc/libvirt/krb5.kqemu /usr/sbin/libvirtd -d --listen In virt-manager I've set both hosts: qemu+tcp://srv1.example.org qemu+tcp://srv2.example.org Since both configs are identical (I've ran diff on them) I am a bit lost at the moment. I do not have any idea why it works for one host, but not the other. Any ideas? -- Thomas

2 1

[libvirt-users] multi-user libvirt / virsh?
by Tomasz Chmielewski 28 Jan '11

28 Jan '11

I was wondering if libvirt / virsh was designed for multi-user environments. I.e. I was thinking of the "roles", i.e. root would be able to do everything, but regular users would only be able to stop/start/suspend, or use vnc with the virtual machines which belong to them. Is it possible with libvirt / virsh? If not, are there any other options to achieve it? -- Tomasz Chmielewski http://wpkg.org

2 1

[libvirt-users] Allocating Sockets Cores and Threads.
by joe fu 27 Jan '11

27 Jan '11

Hi, Is there any other way I could allocate sockets, cores and threads to my virtual machine? I looked over the virsh man page and I didn't find anything that talks about sockets, threads and cores. I edited the vm xml file and added this... <cpu> <topology sockets='1' cores='2' threads='1'/> </cpu> Saw this in the libvirt documents, http://libvirt.org/formatdomain.html Setting this up shows me a single core in my vm, but I don't see any sockets and threads. Can somebody tell me how to do it in a right way so that I see all sockets, cores and threads allocated in my vm. Thanks joe

2 3