[Qemu-devel] difference between 'preallocation=metadata' and 'preallocation=full'

[Kashyap Chamarthy]

Hi,

Is there a difference between the below two invocations of 'qemu-img'
--   I/O performance wise for QCOW2 disk image?

1/ qemu-img create -f qcow2 -o preallocation=metadata
/var/lib/libvirt/images/fedora15.qcow2 8G

2/ qemu-img create -f qcow2 -o preallocation=full
/var/lib/libvirt/images/fedora15.qcow2 8G


According to this patch-->
http://lists.gnu.org/archive/html/qemu-devel/2011-01/msg03017.html

'preallocation=full'  -- which does not only allocate metadata for the
whole image, but also writes zeros to it,
creating a non-sparse image file.


This morning Kevin Wwolf  mentioned on IRC that preallocation is the
best that can be done to extract  near raw performance out of QCOW2.

Can someone please  give insight as to what kind of 'metadata' is
this?  Trying to understand this,  before I try to apply this in my
test lab.


Thanks,
Kashyap Chamarthy