SL6.1 KVM PROBLEM

I am attempting to get KVM working on SL6.1. I have installed KVM and installed an Centos5.6 guest.

The Guest install seemed to go ok but when i try to start the guest i get the following error:

[root@kvm-sl6x ~]# virsh -c qemu:///system start vmcentos56
error: Failed to start domain vmcentos56
error: operation failed: cannot restore domain 'vmcentos56' uuid 65bb2106-5295-5253-2400-f2d0b56836e2 from a file which belongs to domain 'vmcentos56' uuid e6949ba4-86c3-d092-1caf-eee8ec8dc36d

Can anyone assist when this issue.

The guest install script i used is as follows:

virt-install \
-n vmcentos56 \
-r 1024 \
--os-type=linux \
--os-variant=rhel5.4 \
--disk path=/dev/VolGroupGuests/LogVolvmCentos56,device=disk,bus=virtio,sparse=true,format=raw \
-w bridge=br0,model=virtio \
--nographics \
-x console=ttyS0 \
--location http://mirror.ox.ac.uk/sites/mirror.centos.org/5.6/os/x86_64/

The storage is as follows:

[root@kvm-sl6x ~]# pvdisplay
--- Physical volume ---
PV Name /dev/cciss/c0d0p3
VG Name VolGroupGuests
PV Size 435.92 GiB / not usable 3.15 MiB
Allocatable yes
PE Size 4.00 MiB
Total PE 111595
Free PE 106475
Allocated PE 5120
PV UUID Z4ddMr-B9VT-f3CI-uqxS-EZUa-sYJg-QxxXpU

--- Physical volume ---
PV Name /dev/cciss/c0d0p2
VG Name vg_kvmsl6x
PV Size 29.30 GiB / not usable 4.00 MiB
Allocatable yes
PE Size 4.00 MiB
Total PE 7499
Free PE 124
Allocated PE 7375
PV UUID FYH9nI-5L2q-IOo2-verX-6rKr-YA2i-UdNLti

[root@kvm-sl6x ~]# lvdisplay VolGroupGuests
--- Logical volume ---
LV Name /dev/VolGroupGuests/LogVolvmCentos56
VG Name VolGroupGuests
LV UUID fSN3H7-FpOP-tbBD-1yVC-yVYL-AXFq-0PMCcq
LV Write Access read/write
LV Status available
# open 0
LV Size 20.00 GiB
Current LE 5120
Segments 1
Allocation inherit
Read ahead sectors auto
- currently set to 256
Block device 253:2

Any assistance, pointers or suggestions for a possible solution to this issue would be much appreciated.

jk

I have found some SELinux related stuff on chapter 16 of the Red Hat Virtualization guide, that i did not know about.

Having the virtualization host mount guest machine LVM's used to be a disaster with Xen / Xen and SELinux, but it looks as though it might be required now with SELinux and KVM!

I'll read the relevant Red Hat guide before i post next time.

jk

I'm not sure if anyone is interested in this, but i eventually reinstalled the server and got KVM working.

The stuff on Chapter 16 of the Red Hat guide did not seem to be required either it just worked after the second install.

On the second install I created the seperate Physical Vol / Vol Group / LogVol with installer, and unmounted the guest virtual machine's LogVol in /etc/fstab. When i rebooted the server and installed the KVM virtual machine everything working fine.

The difference between the first and second attempt at Scientific Linux and KVM - was that on the first attempt I did not set up the seperate Physical Vol / Vol Group / Log Vol for the virtual machine with the OS installer. I used fdisk, pvcreate, vgcreate, lvcreate. So i must have messed something up at the commend line.

jk

Thanks for the update

I'm still having issues with Scientific Linux 6.1 and KVM.

The reinstall did not do anything to solve this issue. When i tried to install a second virtual machine i got the same error.

I eventually found the time to look into this a bit more over the bank holiday weekend and it seems that i needed to setup an LVM storage pool in accordance with chapter 26 of the Red Hat Virtualization guide.

Once i setup up this storage pool i could install more than one centos vm without any issues. I have not tried any sl5.x or sl6.x installs, but i am working on the assumption that they will not give my any trouble! However, i need to get debian installed as a virtual machine.

I have tried about 6 or 7 installs and whatever adjustments i make i get the same error that i intitally got tried to install the centos vm. Once the install finishes i get the following error:

Requesting system reboot
[ 2263.456698] Restarting system.

Guest installation complete... restarting guest.
ERROR operation failed: cannot restore domain 'vmsqueeze' uuid b1968849-8b0b-5a9a-356d-f5dc2e991062 from a file which belongs to domain 'vmsqueeze' uuid 2b9f511e-c782-44ef-a59e-1e3e037a453c
Domain installation does not appear to have been successful.
If it was, you can restart your domain by running:
virsh --connect qemu:///system start vmsqueeze
otherwise, please restart your installation.
[root@kvm-sl6x ~]#

The script i am using to install the vm is as follows:

virt-install \
--name=vmsqueeze \
--ram=1024 \
--vcpus=1 \
--os-type=linux \
--os-variant=debiansqueeze \
--disk path=/dev/libvirt_lvm/volume3,device=disk,bus=virtio,sparse=true,format=raw \
-w bridge=br0,model=virtio \
--nographics \
-x console=ttyS0 \
--location http://mirror.ox.ac.uk/debian/dists/Debian6.0.2/main/installer-amd64/

I have been fairly maticulous setting up the storage and i am even using the same vg and lv names that are used in the Red Hat Virtualization guide.

[root@kvm-sl6x ~]# virsh pool-list --all
Name State Autostart
-----------------------------------------
guest_images_lvm active yes

[root@kvm-sl6x ~]# lvs
LV VG Attr LSize Origin Snap% Move Log Copy% Convert
volume1 libvirt_lvm -wi-a- 20.00g
volume2 libvirt_lvm -wi-a- 20.00g
volume3 libvirt_lvm -wi-a- 20.00g
LogVolDevShm vg_kvmsl6x -wi-a- 500.00m
LogVolHome vg_kvmsl6x -wi-ao 1.95g
LogVolRoot vg_kvmsl6x -wi-ao 4.88g
LogVolSwap vg_kvmsl6x -wi-ao 6.84g
LogVolTmp vg_kvmsl6x -wi-ao 2.93g
LogVolVar vg_kvmsl6x -wi-ao 4.88g
LogVolVarLog vg_kvmsl6x -wi-ao 6.84g

Does anyone know how to resolve this issue, or is it even possible to get a debian based distro on a SL6 kvm install, as i am now starting to run out of ideas?

jk

I have finally managed to solve this issue with the following commands:

# virsh --connect qemu:///system undefine vmsqueeze
# virsh --connect qemu:///system vol-delete /dev/libvirt_lvm/volume3
# reboot
# virsh --connect qemu:///system vol-create-as guest_images_lvm volume3 20G

I then ran the install script and Squeeze installed.

It seems that full 'virsh --connect qemu:///system' command has to be used to do anything think KVM and SL6.1!

My mistake must have be using 'virsh undefine vmsqueeze' and not deleting and then recreating volume3 with the virsh/libvirt stuff.

I did not have any of this pain when i tested Debian as a KVM host. Everything just seemed to work first time - no virsh storage pool was required (vgcreate etc worked fine) and no need to use virsh to create the LVM's for each virtual machine.

Anyway, KVM is working and hopefully this post might save someone else from having to scratch their head for a couple of days (or perhaps I am just "$%^ at linux!).

jk

If you continue to have issues in VM creation I'd almost suggest a "Desktop" installation, because the GUI VM creation process just works. It eats a little RAM but people like me that type horribly can create a VM in about 30 seconds. Also KVM should be able to host just about any modern x86 OS.

I had a similar issue. I tried changing /etc/libvirt/qemu/<MyVM>.xml's UUID to match the UUID it said it should be, but it doesn't seem to pick it up there. Kindof seems like some stupid file has to be changed to make the UUID match.

virsh managedsave-remove <VMname>

seemed to work for me!

Good work! Thanks for the info

cheers,

QUOTE (serodores @ Nov 16 2011, 04:11 AM)

virsh managedsave-remove <VMname> seemed to work for me!

Also for me.Stupid error.