Hi spxxn,
[QUOTE=spxxn;23023]Long story short, we had a server RAID go down and had to send off for data recovery to get the basic files back. The company sent us the virtual machines that were on this server. I have those virtual machines on a new server and am trying to create a new virtual machine through virt-manager and just mount the disk from the backup. When going through the setup I hit “I have a disk or disk image with installed operating system” , use the default hardware layout, and then click on DISK and select the virtual machine file that I received from data recovery. When I hit OK to try and create all of this I get a “No Kernel Found. Installation may have been unsuccessful” error.
For what it’s worth, I noticed after getting these virtual machines that they were showing that the journal was corrupted. I ran a e2fsck on the server that I’m trying to mount and it’s still not working.[/QUOTE]
I’m not sure I understand that last part - who is reporting the journal problem if you couldn’t start the VM?
What I read from your description is that you had Xen VMs (DomUs) with file-based virtual disks. These files (for the virtual disks) were stored on a RAID. The RAID crashed and was sent in for professional recovery. You wrote that the VMs came back from external recovery, but then mention you try to use the disks “from backup”. You don’t mention if you still have access to the VM definition files (from /etc/xen/vm, if you didn’t change the default location).
So the starting point are the files carrying the virtual disk images, be it from (your) backup or (external) recovery.
Have you tried mounting the file systems of the virtual disks on Dom0 (the “host”)? I would do that before trying to boot up the VMs, just to see what came back from recovery/backup, and to run a full forced fsck on the file systems on those disks. If you have only master files, create copies of the image files first, just in case something goes wrong during fsck or mount!
Once you were able to successfully fsck the FSs on the virtual disks, you could check if your OS files are still available, and where these are on the virtual images. With that information, you can manually adjust the DomU definitions and start the DomU manually. Do you still have the original DomU configuration files? That would make for a good starting point.
If you need or prefer to use virt-manager to create new DomU configurations, I’d suggest to create completely new DomUs (one per OS version), to be used as a template. Then copy the resulting final configuration file and individualize them per DomU you need to recover - that’d be “name”, “uuid”, “disk” and maybe MAC addresses of vifs, if you have these stored in the configuration files. You then also have a chance to adopt to any “unusual” kernel/initrd file configurations you may find inside the DomUs, which prevent virt-manager to successfully boot up the VMs.
Are you familiar with creating/modifying DomU configurations, without using virt-manager? Are you familiar with mounting virtual disk images on the host? I didn’t go into detail here, so please ask if you need assistance with these steps.
Regards,
Jens