Re: Problem with loop mount
On Thu, Jan 18, 2001 at 01:12:06PM -0800, Joey Hess wrote:
> Jon Forsberg wrote:
> > # dd bs=1M seek=300 count=0 of=bigfile
>
> That makes a file that consits of a single 300mb hole. (run du on it)
>
> > If I create the loopfile with "dd bs=1M count=300 if=/dev/zero of=bigfile"
> > instead (which is the "normal" way to do it, I guess) everything works.
> >
> > So is this
> > A. A known problem/limitation in the kernel (can't use loopfiles with holes(?))
>
> I can't find any real documentation about loopback files and whether
> they should support holes. However, the D state means the kernel is in
> an inunteruptable sleep; presumably trying to read the hole.
Ok, but it shouldn't hang totaly in any case, right?
--zzed
Reply to: