Last updated on November 14, 2020 by Dan Nanni
qcow2-type disk image which is used by one my guest VMs on QEMU/KVM. I want to modify the content of the disk image without powering on the VM, and for this I need to mount the disk image somewhere. Is there a way to mount a
qcow2disk image on Linux?
When you run a guest virtual machine (VM) on a hypervisor, you create one or more disk images dedicated to the VM. As a virtual disk volume, a disk image represents the content and structure of a storage device (e.g., a hard disk drive or flash drive) attached to the VM. If you want to modify files in a VM's disk image without powering on the VM, you could mount the disk image. You then would be able to modify the content of the disk image before unmounting it.
In Linux, there are ways to mount a disk image, and different types of disk images require different approaches. If you are using
qcow2-type disk images (used by QEMU/KVM), there are at least two ways to mount them on Linux.
The first method to mount a
qcow2 disk image is by using
libguestfs which provides (among other things) a suite of tools for accessing and editing VM disk images.
libguestfs supports pretty much all kinds of disk images including
qcow2. You can install the
libguestfs toolset on Linux as follows.
$ sudo apt-get install libguestfs-tools
$ sudo yum install libguestfs-tools
libguestfs toolset is installed, you can mount a
qcow2 image using a command-line tool called
guestmount as follows. Note that when a VM is running, you must not mount its disk image in
read-write mode. Otherwise, you run the risk of damaging the disk image. Thus, it is safe to always turn off a VM before mounting its disk image.
$ sudo guestmount -a /path/to/qcow2/image -m <device> /path/to/mount/point
-m <device> option is used to specify a partition within the disk image, that you want to mount (e.g.,
/dev/sda1). If you are not sure what partitions exist in the disk image, you can supply any bogus device name. The
guestmount tool will then show you available device names to choose from.
For example, to mount
/dev/sda1 in the disk image
/mnt mount point, run the following.
$ sudo guestmount -a /var/lib/libvirt/images/xenserver.qcow2 -m /dev/sda1 /mnt
By default, the disk image is mounted in
read-write mode. So you can modify any file in
/mnt after mounting it.
If you want to mount it in
$ sudo guestmount -a /var/lib/libvirt/images/xenserver.qcow2 -m /dev/sda1 --ro /mnt
To unmount it:
$ sudo guestunmount /mnt
Another method to mount a
qcow2 disk image is via
qemu-nbd, a command-line tool that exports a disk image as a network block device (
You can install
qemu-nbd on Linux as follows.
$ sudo apt-get install qemu-utils
$ sudo yum install qemu-img
To mount a
qcow2 disk image, first export the image to
nbd as follows.
$ sudo modprobe nbd max_part=8 $ sudo qemu-nbd --connect=/dev/nbd0 /path/to/qcow2/image
The first command loads the
nbd kernel module. The
max_part=N option specifies the maximum number of partitions we want to manage with
nbd. The second command exports the specified disk image as a network block device (
/dev/nbd0). As a network block device, you can use
/dev/nbd2, etc. whichever is unused. As for the disk image, make sure to specify its full path.
For example, to export the image
$ sudo qemu-nbd --connect=/dev/nbd0 /var/lib/libvirt/images/xenserver.qcow2
After this, existing disk partitions in the disk image will be mapped to
To check the list of
nbd-mapped partitions, use
$ sudo fdisk /dev/nbd0 -l
Finally, choose any one partition (e.g.,
/dev/nbd0p1) and mount it to a local mount point (e.g.,
$ sudo mount /dev/nbd0p1 /mnt
Now you will be able to access and modify the content of the disk image's mounted partition via
/mnt mount point.
Once you are done, unmount it, and disconnect the disk image as follows.
$ sudo umount /mnt $ sudo qemu-nbd --disconnect /dev/nbd0
Please note that this article is published by Xmodulo.com under a Creative Commons Attribution-ShareAlike 3.0 Unported License. If you would like to use the whole or any part of this article, you need to cite this web page at Xmodulo.com as the original source.