I recently replaced my motherboard and, when attempting to boot, I was dropped into the recovery shell because it was unable to find the partitions by UUID. The recovery console could not see any hard disks; there were no /dev/sd* devices and no /dev/disk structure. I booted into a live USB and regenerated the initramfs, which solved the problem. That episode is chronicled over at the Google+ community: https://plus.google.com/107119852442751 … 7PTjFc9N3c.
Now that I am back up and running, I attempted to start one of my virtual machines (using qemu-kvm) only to encounter the same issue: can't find disks by UUID, dropped into recovery shell, no disks visible. I then attempted to follow the same procedure as above, but this time the live USB can't see any of the devices either (other than itself as /dev/sda). I've tried with qemu-system-x86_64 and virt-manager/libvirtd, but they produce the same result.
Here's the qemu line:
sudo QEMU_AUDIO_DRV=alsa /usr/bin/qemu-system-x86_64 -name vmname -enable-kvm \ -m 4096 -smp 2 \ -usb -usbdevice host:090c:1000 \ -drive file=/dev/mapper/drive_new,format=raw \ -device AC97 \ -net nic -net bridge,br=br0 \ -boot menu=on
No errors are reported in either the terminal or the libvirt logs, but the machine does report this on boot, prior to dropping me into the recovery shell:
[...] ata_piix 0000:000:01.1 device not available (can't reserve [io ...]) [...] pata_acpi 0000:000:01.1 device not available (can't reserve [io ...])
I tried also booting into an old F17 live disk I had as well, but encountered the exact same problem. Considering the problem occurs within and without the VM and that it occurs in Fedora too, I'm pretty sure there's something going on with the motherboard, but I am not a hardware expert and am not sure where to look. I explored the BIOS and didn't see anything obvious, though tried a few settings to see if they changed behavior; they did not and were reset. The new motherboard is an MSI 970A-G43.
Any help anyone can provide would be most appreciated. Thanks!
Not sure if this will help, but I had the same issue as you. Everything was working fine and then I'm not sure if a `seabios` or `qemu` upgrade borked everything and I had similar errors to you.
Turns out my mirror was a little behind and there were more recent seabios and qemu updates. I pulled from another mirror and updated and everyting worked again.
qemu 1.5.2-1 didnt work while version (1.5.3-1) coupled with the latest seabios update (not sure what version) did.
Give the update a shot, and let me know how you get on