I'm running Debian Squeeze and kvm/qemu which installed successfully. Is there any equivalent to the way xen did it, where it bootstraps in and you just console in and set the rest up, or do you have to build vnc somehow to finish the guest install? I don't want a gui on either host or guest. I have tried to start a new vm using: Code: virt-install --connect qemu:///system -n vm11 -r 1024 --vcpus=8 -f /home/vm/vm11.qcow3 -s 100 -c /some/path/debian/debian-testing-amd64-CD-1.iso -vnc :1 --noautoconsole --os-type linux --os-variant debianSqueeze --accelerate --network=bridge:br0 --hvm which was successful, then try to connect to it with Code: vncviewer :1 or vncviewer my.lo.cal.ip:1 but I get Code: Error: can't open display I have also tried connecting to it with a Debian server running a gui on the same subnet using vnc, but it won't connect. Then I tried to use virt-install with the -serial switch, but that just errored out and wouldn't build the machine
Ubuntu has the vmbuilder tool which does it, but I don't know if it's available for Squeeze as well. Otherwise, you must use VNC, but you don't need to install a desktop on the host or guest. You jsut need some desktop system with a VNC client.
vmbuilder is available for Etch, but still waiting to be ported to Squeeze, it errors out when I attempt to run it. on a Debian machine with a gui and vnc on the same subnet trying to connect to ip.of.kvm.box:1 I get a connection refused, though I can vnc/rdp using this same client to several other boxes on the subnet, do I need to install any other app on the KVM box? The KVM box doesn't have a gui. If I can't get this to work, can I do the same thing in virtualbox, and is virtualbox performance similar to KVM? Most of the time I don't use gui's at all, that's a real disadvantage for server admin, which is why I used xen until it got pulled from the kernel in Squeeze
Is there a firewall on the host? What's the output of Code: iptables -L ? Yes, you can do this with vBoxheadless (see the tutorials on this site), but you still need a remote desktop connection. I can't say much about the performance, but I guess KVM should be more performant since it's an official part of the kernel.
Since VirtualBox still needs a vnc server, I decided to just work on KVM again, in the meantime I got vncviewer to run a remote box after installing (on server) Code: apt-get install tightvncserver xtightvncviewer vncserver :2 -geometry 1024x768 -depth 16 -pixelformat rgb565 now, after attempting to kill all kvm vm's from the original attempt (before trying VirtualBox) and then trying to install a new one with my :2 display selected, like this: Code: virt-install --connect qemu:///system -n sq1 -r 1024 --vcpus=8 \ -f /home/vm/sq1.qcow5 -s 50 \ -c /path/to/debian-testing-amd64-CD-1.iso \ -vnc :2 --noautoconsole --os-type linux --os-variant debianSqueeze \ --accelerate --network=bridge:br0 --hvm I get Code: ERROR Guest name 'c' is already in use. Why do I get this, I'm not specifying guest 'c'? When I enter virsh and run Code: virsh # list --all Id Name State ---------------------------------- - c shut off - vm10 shut off virsh # destroy c error: Failed to destroy domain c error: Requested operation is not valid: domain is not running virsh # domstate c shut off virsh # dominfo c Id: - Name: c UUID: b341252f-46f2-ff62-ef38-71ef2d441cf0 OS Type: hvm State: shut off CPU(s): 8 Max memory: 1048576 kB Used memory: 1048576 kB Autostart: disable Security model: qemuDAC Security DOI: how do I remove that guest and why can't I create a new one with the above virt-install command? I tried to start 'c' vm and then looked at it, here's what I have: Code: ps aux | grep virt root 1932 0.0 0.0 245148 8452 ? Sl 08:33 0:00 /usr/sbin/libvirtd -d 106 2449 0.1 0.0 1732084 11884 ? S 09:27 0:02 /usr/bin/qemu-system-x86_64 -S -M pc-0.11 -m 1024 -smp 8 -name c -uuid b341252f-46f2-ff62-ef38-71ef2d441cf0 -nographic -monitor unix:/var/lib/libvirt/qemu/c.monitor,server,nowait -boot c -drive file=/home/vm/vm11.qcow3,if=ide,bus=0,unit=0,format=raw -drive if=ide,media=cdrom,bus=1,unit=0 -net nic,macaddr=52:54:00:37:4f:e1,vlan=0,name=net0 -net tap,fd=46,vlan=0,name=hostnet0 -serial pty -parallel none -usb -usbdevice tablet -vga none -balloon virtio but it's not connected to a display so I can't view it on a remote machine, and even if I Code: kill 2449 I still get the same error when trying to create a new hostname. I think the install is hung, I just can't figure out how to unhang it
I have also tried to Code: virsh unregister c but it said there's no such domain, then I updated my switches to virt-install man page for squeeze, so it looks like: Code: virt-install --connect qemu:///system -n sq1 -r 1024 --vcpus=8 \ --file=/home/vm/sq1.qcow5 size=50 \ --location=http://ftp.us.debian.org/debian/dists/squeeze/main/installer-amd64/ \ -vnc :2 --noautoconsole --os-type=linux --os-variant=debiansqueeze \ --accelerate --network=bridge:br0 --hvm but I get the same error: Code: ERROR Guest name 'c' is already in use. I have also tried using Code: virt-install --prompt to do it interactively, but that script errors out with Code: ERROR this function is not supported by the hypervisor: virConnectNumOfDefinedStoragePools Traceback (most recent call last): File "/usr/bin/virt-install", line 972, in <module> main() File "/usr/bin/virt-install", line 762, in main options.sparse, options.nodisks, guest, conn) File "/usr/bin/virt-install", line 314, in get_disks get_disk(disk, size, sparse, guest, conn, is_file_path) File "/usr/bin/virt-install", line 280, in get_disk d = cli.disk_prompt(None, kwargs) File "/usr/lib/pymodules/python2.5/virtinst/cli.py", line 427, in disk_prompt dev = VirtualDisk(**arg_dict) File "/usr/lib/pymodules/python2.5/virtinst/VirtualDisk.py", line 372, in __init__ self.__validate_params() File "/usr/lib/pymodules/python2.5/virtinst/VirtualDisk.py", line 863, in __validate_params self.__check_if_path_managed() File "/usr/lib/pymodules/python2.5/virtinst/VirtualDisk.py", line 733, in __check_if_path_managed os.path.dirname(self.path)) File "/usr/lib/pymodules/python2.5/virtinst/util.py", line 553, in lookup_pool_by_path inactive_list = conn.listDefinedStoragePools() File "/usr/lib/python2.5/site-packages/libvirt.py", line 1583, in listDefinedStoragePools if ret is None: raise libvirtError ('virConnectListDefinedStoragePools() failed', conn=self) libvirtError: this function is not supported by the hypervisor: virConnectNumOfDefinedStoragePools after I specify filesystem type. Racking my brains out
Code: virsh # undefine c error: failed to get domain 'c' error: Unknown failure virsh # list --all Id Name State ---------------------------------- virsh # quit #: virt-install --connect qemu:///system -n sq1 -r 1024 --vcpus=8 \ > --file=/home/vm/sq1.qcow5 size=50 \ > --location=http://ftp.us.debian.org/debian/dists/squeeze/main/installer-amd64/ \ > -vnc :2 --noautoconsole --os-type=linux --os-variant=debiansqueeze \ > --accelerate --network=bridge:br0 --hvm ERROR Guest name 'c' is already in use.
I submitted a bug report to Debian Squeeze bugtracker, will post back here if I hear something, unless someone here has any suggestions left?
ok, even after you run virsh, you still have to connect to the local machine to undefine it: Code: virsh -c qemu:///system undefine c but then something is wrong with my setup I'm using the create a new domain with virt-install, this tries to create another domain named 'c' for some reason, I'm betting it's due to a deprecated switch of some kind, possibly on the .iso source location, so I'm trying: Code: virt-install --connect qemu:///system --name sq3 --ram 1024 --vcpus=8 -f /home/vm/vm14.qcow5 -s 100 --location=http://ftp.us.debian.org/debian/dists/squeeze/main/installer-amd64/ --vnc :2 --network=bridge:br0 --hvm but that gives me the error: Code: ERROR operation failed: failed to retrieve chardev info in qemu with 'info chardev' I'll post if I get it working. In the meantime, just use Debian Lenny, it's far easier, I needed Squeeze because Lenny didn't have my NIC drivers for a HP DL180 G6 and I didn't want to compile them, they are part of Squeeze though.
Did you ever get this working? I'm having the same issue in Ubuntu Lucid when tring to start a KVM VM with /dev/mapper/nvidia_ebgdfiad1 as the hard disk.
Same Error on Ubuntu-Server-10.10 Here is what I used: virt-install --connect qemu:///system -n ubuntu-vm1 -r 2048 --vcpus=4 -f /root/vmimages/kvm1.qcow2 -s 12 --cpuset=4,5,6,7 -c /root/vmimages/ubuntu-10.10-server-amd64.iso --os-type linux --network bridge=virbr0 --accelerate --virt-type kvm --hvm Here's what I get: Starting install... ERROR operation failed: failed to retrieve chardev info in qemu with 'info chardev' Domain installation does not appear to have been successful. If it was, you can restart your domain by running 'virsh start ubuntu-vm1'; otherwise, please restart your installation. ERROR operation failed: failed to retrieve chardev info in qemu with 'info chardev' Traceback (most recent call last): File "/usr/bin/virt-install", line 1033, in <module> main() File "/usr/bin/virt-install", line 915, in main start_time, guest.start_install) File "/usr/bin/virt-install", line 957, in do_install dom = install_func(conscb, progresscb, wait=(not wait)) File "/usr/lib/pymodules/python2.6/virtinst/Guest.py", line 973, in start_install return self._do_install(consolecb, meter, removeOld, wait) File "/usr/lib/pymodules/python2.6/virtinst/Guest.py", line 1038, in _do_install "install") File "/usr/lib/pymodules/python2.6/virtinst/Guest.py", line 1009, in _create_guest dom = self.conn.createLinux(start_xml, 0) File "/usr/lib/python2.6/dist-packages/libvirt.py", line 1277, in createLinux if ret is None:raise libvirtError('virDomainCreateLinux() failed', conn=self) libvirtError: operation failed: failed to retrieve chardev info in qemu with 'info chardev' If someone has found a way around it please let me know. - Rohan
i'm using also squeeze/amd64, but have no virbr0 interface at all. On virsh i run net-start default, having: error: Failed to start network default error: internal error '/sbin/iptables --table filter --delete INPUT --in-interface virbr0 --protocol udp --destination-port 69 --jump ACCEPT' exited with non-zero status 1 and signal 0: iptables: Bad rule (does a matching rule exist in that chain?). net-edit default gives: <network> <name>default</name> <uuid>4b8fa970-631c-9313-e639-209a482a45a4</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> any hints ? regards, Pablo
You might want to try this tutorial: http://www.howtoforge.com/virtualization-with-kvm-on-a-debian-squeeze-server
