Am 27.05.2014 um 19:45 hat Benoît Canet geschrieben: > The Tuesday 27 May 2014 à 17:29:16 (+0200), Paolo Bonzini wrote : > > Il 27/05/2014 18:00, Benoît Canet ha scritto: > > >> Are you worried that the host file system will lay out data poorly > > >> because the file looks like this? > > >> > > >> | header (1C) | refcounts (2C) | L1 (512B) | hole | Next cluster | > > >> > > >> B = bytes > > >> C = clusters > > >> > > >> My guess is the next cluster will be aligned to a reasonable boundary on > > >> the physical disk. > > > > > >I have some kind of doubt. Does anyone knows a filesystem guru ? > > > > Not a guru, but indeed there is a risk that the layout will be worse > > than necessary. > > > > I think holes are ignored unless they are big enough, but 64K-512 is > > probably enough to create one. Indeed here I get this: > > > > $ qemu-img create -f qcow2 foo.qcow2 10G > > $ qemu-io -c 'write 0 512' foo.qcow2 > > $ qemu-img map -f raw foo.qcow2 10G > > Offset Length Mapped to File > > 0 0x31000 0 foo.qcow2 > > 0x40000 0x20000 0x40000 foo.qcow2 > > That seems to be 4k aligned. It's better than I though.
These 4k are probably the host file system block size. Kevin
