[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Importing KVM ivda drives as xdva drives

On Sat, Jun 9, 2018 at 2:06 AM, Predrag Punosevac <punosevac72@xxxxxxxxx> wrote:
> Hi Guys,
> This is my first post to this mailing list so please excuse my
> ignorance. I started running Alpine Linux Dom0 about half a year ago and
> I am very happy with it.
> Recently I inherited a KVM hypervisor with about dozen guests some of
> which are Windows servers.  All KVM guests are either raw or qcow2 image
> file formats. I am trying to move them from KVM host to Dom0 (not
> block devices or block device partitions).
> I did a bit of playing trying to migrate RHEL 7.5 image to Xen Dom0
> Alpine. It seems that the only obstacle in importing KVM images into the
> Xen is the disk naming. Namely KVM is using virtio (paravirtualization
> driver) for disks which are named as
> /dev/vda
> for a OS drive. Xen on the another hand is using different naming
> convention where the drives inside disk images .img are denoted for
> example as
> /dev/xvda
> When I create DomU out of KVM image file the both process proceed
> successfully until it drops into dracut shell as it can't find root
> partition which is on the /dev/vda drive. Is the HDDs naming superficial
> problem or is it fundamental incompatibility on the level of disk
> drivers? This is my rhel.cfg file

It would be helpful to have the actual console output from the guest;
look for the 'serial' option in the xl.cfg manpage.

But the first thing to check is whether your grub and fstab entries
are using persistent device naming[1], or are using hard-coded device
names.  If your guest grub entry says "root=/dev/vda", then Linux will
look for that specific device, not xvda, because that's what you've
told it to do; similarly with secondary entries in fstab.

If you do have hard-coded device names, the quick-and-dirty option is
to go around change vdN to xvdN in grub and fstab.  The more robust
long-term option is to use UUIDs (or something else); that should be
robust against any future device renames as well (for example, if you
wanted to try moving back to KVM).


[1] https://wiki.archlinux.org/index.php/persistent_block_device_naming

Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.